Skip to main content

Microresonator based Frequency Comb Generators

Final Report Summary - µCOMB (Microresonator based Frequency Comb Generators)

--Project context
The development of frequency combs based on mode locked fiber or Titanium Sapphire lasers in the late 1990s has paved the path for high precision spectroscopy of optical frequencies, coherently linking optical frequencies with the radio- or microwave frequency domain. Frequency combs are therefore essential in optical frequency metrology and became valuable tools in other application areas, including attosecond pulse generation, astrophysical spectrometer calibration, optical waveform synthesis or molecular spectroscopy. An entirely different and original approach based on the use of optical micro cavities was demonstrated by the research group of Prof. T. J. Kippenberg: In ultra-high Q micro resonators, light can be confined in a small volume over extended amounts of time. There, the interaction of a strong CW pump laser with the resonator modes gives rise to a broad comb-like emission spectrum. The process underlying the formation of the frequency comb exploits the nonlinearity of glass itself: due to the intensity dependent refractive index (Kerr nonlinearity) two pump photons can be annihilated and give rise to a frequency up and down shifted pair of photons. This process can cascade and thereby lead to the formation of a broad comb, whereby the energy conserving nature of parametric process ensures equidistance of all teeth of the comb. This presented a departure from the conventional dogma that optical frequency combs can only be generated by mode locked lasers. Instead, optical nonlinearity provides the same capability, albeit with several distinct advantages:

First, the novel – monolithic – approach promises unprecedented compactness, direct fiber optic integration and high efficiency. Even more important, the comb provides access to either unattainable high repetition rates (>40 GHz).

The potential applications in both academia and in industry of this novel type of monolithic combs are diverse and a few examples are outlined below:
Telecommunication: The mode spacing in the 100 GHz range along with high power (milli-Watt per mode) make them promising for application in telecommunications. The ITU grid spacing is 100 GHz and therefore makes this source very attractive as channel generator, replacing a bank of conventional DFB lasers.
Astrophysical spectrometer calibration: On the scientific side, the demonstrated microwave mode spacing is a key prerequisite for the calibration of astrophysical Echelle spectrometers used for planet searches. So far, the low repetition rate has precluded simple femtosecond laser based combs from being used for this application but required filtering out undesired modes. In contrast, the monolithic comb generators can emit 20 GHz combs, ideally suited for astrophysical spectrometer calibration.
Low phase noise microwave generators: The demonstrated four wave mixing process also enables to create low phase noise microwave signals with applications in areas such as Radar technology.

The key advantages of monolithic frequency comb generators are:
• Unprecedented form factor (milli-meter size, gramm weight)
• Repetition rates >10 GHz, which so far have been inaccessible but are pivotal for many applications.
• High efficiency, low power consumption due to ultra low loss optical modes
• High power per comb mode (1 mW typical)
• No free space optics or piezo elements required for controlling offset frequency and repetition rate
• The monolithic frequency comb generation uses parametric gain and does not require mode locking conditions to be satisfied; large insensitivity to external perturbations
• Octave spanning spectra without the need for any further external spectral broadening.
• Monolithic and extremely compact design; not prone to vibrations or acceleration.

Thus, micro resonator Kerr combs offer unique possibilities in an attractive physics package. Of course, this has already been adapted and stimulated research in other laboratories worldwide. Indeed, shortly after our report on the first Kerr-comb appeared, a group in the USA at the NASA Jet Propulsion Laboratory was able to demonstrate the same generation process in crystalline microresonators of CaF2. This material exhibits transparency from 160 nm up to ca. 8 micron, which could therefore enable to significantly extend the frequency range of frequency combs to the mid IR and are part of the consortiums research efforts. In 2009, in the work of another US based research group, the Kerr-comb formation could also be observed in microrings which were fabricated in a fully CMOS compatible process in Silicon Nitride chips, boosting the integration of frequency comb technology in photonic chip architectures. Up to now about ten research groups participate to the field of Kerr comb research, working on a multitude of host materials and fabrication techniques, showing the importance and interest in this new technology.

The objective of this research project is to turn over the monolithic frequency comb generators into a commercially viable product and to address several key scientific and technological challenges which are outstanding. In this context it is envisioned to pursue several interesting topics and applications. These are fully self-referenced optical frequency combs in the telecommunication wavelength range. Second, a novel class of crystalline microresonators is to be investigated in views of creating optical frequency combs in the visible and mid-infrared wavelength range. Third, a proof of concept study will be initiated if crystalline microresonators can serve as compact optical reference cavities. Finally, the performance of these comb generators will be investigated in an application. The development of product integration and packaging strategies as well as the necessary exchange of knowledge and technology between the academic and the industrial partners is especially important and another prime objective.

--Summary of main achievements
Comparing the original project proposal’s goals with the achievements demonstrated within this collaboration, several key milestones remain outstanding, which reflects the inherent difficulty to predict the outcome of fundamental research over such a long period. However, our collaborative project has taken many necessary and important steps for realizing such a device, many possible applications of Kerr combs and high Q resonators were investigated, and the dream of a fully self-referenced microresonator based comb is in very close reach:
--Towards self referenced Kerr combs
A crucial task for obtaining a fully stabilized and self-referenced frequency comb is the generation of octave spanning spectra with low phase noise. Important figures of merit for this are a precise knowledge about the dispersion, its measurement and engineering, knowledge about the comb formation process and the search for low phase noise states of Kerr combs. Significant steps in all of these areas were achieved during this project:
By controlling and optimizing the dispersion of silica microtoroids, we successfully demonstrated for the first time an octave spanning Kerr comb spectrum, covering a wavelength range from 990nm to 2170 with a large mode spacing of 850GHz and tunability across a full FSR. However, large phase noise made self referencing impossible in this platform.
The crucial element for understanding the physical process of comb formation is the dispersion, with two competing contributions from material and geometric properties. Mutually cancelling these contributions allows the realization of very small dispersion values, unamenable to experimental measurements so far. We therefore developed a novel method by which the narrow resonance of optical microresonators can be accurately and rapidly measured. The technique being realized before the start of this program, its knowledge was successfully transferred between EPFL and Menlo and is now the standard tool for dispersion characterization at both sites.
We explored and refined resonator fabrication techniques and are now able to create Kerr combs in various structures and host materials. Among these are the promising techniques of polished crystals with record Q factors exceeding 108. This type of resonators is readily produced based on well controlled mechanical grinding and polishing processes at EPFL, a process which was successfully transferred to Menlo throughout this IAPP collaboration. In addition, the very promising platform of lithographically defined microrings was established at EPFL. These fully CMOS compatible devices allow for the integration of resonators and coupling waveguides in a monolithic structure, promoting the integration of Kerr comb generators in integrated microphotonic and -electronic devices. With this platform, we are additionally able to engineer the dispersion of integrated silicon nitride based ring resonators through conformal coating via atomic layer deposition. Both, magnitude and bandwidth of anomalous dispersion can be significantly increased with this process.
A severe limitation in Kerr combs is phase noise, observed in the form of linewidth broadening, multiple repetition-rate beat notes and loss of temporal coherence. These phenomena are not explained by the current theory of Kerr comb formation, yet understanding them is crucial to the maturation of Kerr comb technology. The above techniques of dispersion measurement applied to crystalline MgF2 and planar Si3N4 microresonators, allowed us to reveal the universal, platform-independent dynamics of Kerr comb formation, enabling the explanation of a wide range of phenomena not previously understood, as well as identifying the condition for, and transition to, low-phase-noise performance.
A breakthrough achievement was the realization of modelocked states in microresonators: the formation of dispersive solitons in crystalline resonators was for the first time ever demonstrated in MgF2 resonators, providing a pulsed output combining the favorable properties of short pulses (200fs) and high repetition rate (~15 and 40 GHz). These low noise states finally pave the way for fully stabilized and self-referenced Kerr combs after external broadening in highly nonlinear fibers. Only a few days after the end of this project, coherence of a 2/3 octave spectrum was demonstrated, the very first demonstration of a fully locked and self-referenced Kerr comb is finally within reach. Additionaly, solitons were also observed with our SiN chip based platform, where dispersion management enabled a 2/3 octave output without external broadening.
--Extension of Kerr comb spectra to visible and Mid IR regions.
Research activity was also dedicated to the extension of the frequency comb technology to the mid-IR region. Accessing the mid-infrared “molecular fingerprint” region (2-20 µm) is of great interest for spectroscopy as most of the strong fundamental molecular vibration frequencies are located in this range. So far only a limited number of frequency comb sources produced have been realized. We demonstrated, for the first time, mid-infrared frequency comb generation from monolithic optical whispering-gallery mode microresonators principally allowing for compact, broadband and widely spaced combs with high power per comb line.
--Investigation of crystalline resonators as optical frequency references
Relying on ultra high Q crystalline resonators, we explored for the first time the use of these resonators as stable flywheels for optical atomic clocks. We have examined the potential of crystalline whispering gallery mode (WGM) resonators as reference cavities by measuring the frequency stability of an external cavity laser locked to a WGM and thereby demonstrated a linewidth narrowing to below 290 Hz, attaining the limit imposed by thermodynamic fluctuations.
--Demonstration of high speed telecom data transmission
Relying on the new integrated SiN chip platform, we proved the ability of the generated frequency combs to transmit data at high rates in a joint experiment with the Karlsruhe Institute of Technology. The aggregated transmission speed over 20 Kerr comb channels was 1.4 Tbit/s with a low bit error rate. This was only possible using advanced modulation formats and high modulation frequencies relying on the low noise of the frequency comb.
--Towards product integration and application tests
The necessary infrastructure for coupling, pumping and characterization of the resonators was set up at Menlo Systems. This includes a high performance optical frequency reference, capable to emit optical low phase noise radiation with a linewidth below 1 Hz, a novel beat-detection scheme for advanced comb calibrated dispersion measurement with conventional ECDLs. In addition key experimental core techniques have been transferred from EPFL to Menlo. First, crystalline resonators are ground and polished on a high precision polishing rig now also at Menlo. Large Q factors, FWM spectra and signatures of soliton states have been observed since. Second, a fiber tapering rig was built based on EPFL experience also at Menlo. Here, tapers necessary for coupling to crystalline resonators are regularly produced. In addition, the process was also transferred to tapering photonic crystal fibers for demanding spectral broadening applications.
Building on the results from SiN resonator fabrication, a transportable bread board demonstrator has been developed at Menlo Systems, comprising a silicon nitride microring, coupled with an on-chip waveguide and tapered fibers. An external cavity diode laser and two fiber amplifiers provide the necessary power and control to characterize the resonators and to create broad spectra (1325 – 1945 nm) with large mode spacing (200GHz). The setup is compact and transportable as evidenced by its first presentation to the public at the CLEO 2012 exhibition (San Jose, USA) (Fig. 2). Investigations on various packaging strategies for advanced fiber-chip coupling, sufficing the extreme conditions high power handling and sub µm alignment and drift stability will further decrease the system size and allow for integration within an overall compact device.
The same goal is addressed in the development of a first demonstrator of a crystalline resonator based low noise Kerr comb. A stable taper-crystal packaging, compact seed sources and drive electronics are currently integrated into a rack sized device, scheduled for autumn 2014.
These important steps bring the realization and characterization of the first self-referenced and fully stabilized Kerr comb in reach, combining all favorable properties of compactness and efficiency while yielding broad optical spectra with conventionally inaccessible large repetition rates >10 GHz.

--Main Publications
The scientific progress led to various publications in renowned peer reviewed journals
[1] P. Del’Haye, T. Herr, E. Gavartin, M. Gorodetsky, R. Holzwarth, and T. Kippenberg, “Octave Spanning Tunable Frequency Comb from a Microresonator,” Physical Review Letters 107, 1–4 (2011)
[2] T. Herr, K. Hartinger, J. Riemensberger, C. Y. Wang, E. Gavartin, R. Holzwarth, M. L. Gorodetsky, and T. J. Kippenberg, “Universal formation dynamics and noise of Kerr-frequency combs in microresonators,” Nature Photonics 6, 480–487 (2012)
[3] J. Riemensberger, K. Hartinger, T. Herr, V. Brasch, R. Holzwarth, and T. J. Kippenberg, “Dispersion Engineering and Measurement of Silicon Nitride based Ring Resonators coated by Atomic Layer Deposition,” in preparation (2012).
[4] T. J. Kippenberg, R. Holzwarth, and S. A. Diddams, “Microresonator-based optical frequency combs.,” Science 332, 555–559 (2011)
[5] C. Y. Wang, T. Herr, P. Del’Haye, A. Schliesser, J. Hofer, N. Picqué, and T. J. Kippenberg, “Mid-Infrared Optical Frequency Combs based on Crystalline Microresonators,” arxiv 1109.2716 1–6.
[6] A. Schliesser, N. Picqué, and T. W. Hänsch, “Mid-infrared frequency combs,” Nature Photonics 6, 440 - 449 (2012)
[7] J. Alnis, a. Schliesser, C. Wang, J. Hofer, T. Kippenberg, and T. Hänsch, “Thermal-noise-limited crystalline whispering-gallery-mode resonator for laser stabilization,” Physical Review A 84, 2–5 (2011)
[8] I. Fescenko,J. Alnis,A. Schliesser, C. Y. Wang,T. J. Kippenberg, T. W. Hänsch, “Dual-mode temperature compensation technique for laser stabilization to a crystalline whispering gallery mode resonator,” Optics Express 20/17, 19185-19193 (2012)
[9] T. Herr, V. Brasch, J. D. Jost, C. Y. Wang, N. M. Kondratiev, M. L. Gorodetsky, T. J. Kippenberg, “Temporal solitons in optical microresonators,” Nature Photonics, 8, 145-152 (2014)
[10] T.Herr V. Brasch, J.D. Jost, I. Mirgorodskiy, G. Lihachev, M.L. Gorodetsky, T.J. Kippenberg, “Dispersion and soliton formation in microresonators,” Physical Review Letters (in review)
[11] Joerg Pfeifle , Victor Brasch , Matthias Lauermann , Yimin Yu , Daniel Wegner , Tobias Herr , Klaus Hartinger , Philipp Schindler , Jingshi Li , David Hillerkuss , Rene Schmogrow , Claudius Weimann , Ronald Holzwarth , Wolfgang Freude , Juerg Leuthold , Tobias J. Kippenberg , Christian Koos, “Coherent terabit communications with microresonator Kerr frequency combs,” Nature Photonics, 8, 375-380 2014)