High-speed communication systems that use optical fibres often require hundreds of lasers. An approach that replaces these lasers with a single, ring-shaped optical device offers many technical advantages. See Letter p.274
Optical-fibre communication systems form the backbone of the Internet. Current systems rely on a technology called wavelength-division multiplexing to transmit digital information at high speeds. On the transmitter side, this technology combines (multiplexes) many optical channels into a single optical fibre. Each channel uses a laser of a different frequency, and hundreds of lasers are typically needed to occupy the bandwidth available in a fibre-optic link. On page 274, Marin-Palomo et al.1 demonstrate that all of these lasers can be replaced by a single light source known as a microresonator frequency comb — a development that could lead to extremely fast data transmission.
A microresonator frequency comb is an optical device that allows light of many optical frequencies to be generated in a micrometre-scale platform (Fig. 1). Tobias Kippenberg, one of the current paper's co-authors, helped to pioneer this technology about a decade ago2. The device essentially consists of a light source, called a pump laser, and a microresonator — a set-up also known as an optical cavity, which is used to trap light at certain 'resonance' frequencies. The frequency of the pump laser is closely tuned to a particular resonance of the cavity, and for microscale low-loss cavities, the light is highly confined. The authors made their cavity from a nonlinear material, which allowed the photons from the pump laser to be converted into photons of different frequencies2.
Under the right conditions, the new optical frequencies are phase-locked. This means that at certain times there is constructive interference between all the frequencies (the crests and troughs of the light waves reinforce each other), leading to a substantial build-up of optical power inside the cavity. The resulting waveform consists of a sequence of pulses known as dissipative Kerr solitons. The formation of these solitons in an optical cavity requires a fine balance between the properties of the cavity and the pulses themselves3.
Although Marin-Palomo et al. are not the first to observe dissipative Kerr solitons in optical cavities4, they are the first to use these light sources for optical communications. The authors manufactured their optical cavity using advanced microlithographic techniques. The cavity consists of a microresonator arranged in a ring-like structure (with a radius of 240 micrometres) that is made of silicon nitride, a widely used thermal insulator in the electronics industry. The authors carefully engineered the cavity's geometry to enable the generation of more than 90 optical frequencies from a single pump laser. These frequencies entirely cover two of the bands used for optical-fibre communications (the C- and L-bands), corresponding to a bandwidth of approximately 10 terahertz (1 THz is 1012 Hz). The authors can control the frequency spacing between the channels with a precision of approximately 200 kHz — a feature that, besides its uses in optical-fibre communications, offers prospects for molecular spectroscopy5.
Marin-Palomo and collaborators report a series of impressive system-level demonstrations, whereby the individual channels are multiplexed to yield a data-transmission rate of more than 50 terabits per second. The current transmission-speed record6 is 2,150 terabits per second, but involves a special type of optical fibre and a different kind of laser frequency comb. The key aspect of the authors' microresonator comb is that it achieves an astonishing performance in a microscale platform. With recent developments in 3D photonic integrated circuits7, one can start to dream about combining all of the necessary optoelectronic components of a comb-based wavelength-division-multiplexing system, as required for practical applications.
One concern when generating many frequency components from a single laser is the amount of power that can be obtained per channel. A fundamental drawback with dissipative Kerr solitons is that they have unfavourable power-conversion efficiencies8. In the case of Marin-Palomo and colleagues' system, less than 1% of the laser pump's power is transferred to the newly generated frequencies. There are alternative microresonator combs that have much higher power-conversion efficiencies9, but they have not yet been investigated in the context of optical-fibre communications. Increasing the efficiency of microresonator combs will be essential for future optical-fibre communication systems, which will use a special type of fibre containing multiple spatial channels to achieve unprecedented transmission speeds6.
Marin-Palomo et al. have clearly demonstrated that dissipative Kerr solitons can be used for wavelength-division multiplexing. Using a light source that has phase-locked frequencies represents a fundamental difference from an array of individual lasers because the frequency spacing between channels is fixed. This aspect might be the key to mitigating transmission impairments10 or drastically simplifying the way signals are received. In this respect, the use of laser frequency combs (be it in the form of dissipative Kerr solitons or something else) constitutes a pivotal change in the design of optical-fibre communication systems. Exploiting their unique properties will require a collaborative effort between the disciplines of photonic integration, fibre optics, ultrafast optics, computer engineering, information theory and signal processing.Footnote 1
Marin-Palomo, P. et al. Nature 546, 274–279 (2017).
Del'Haye, P. et al. Nature 450, 1214–1217 (2007).
Herr, T. et al. Nature Photon. 8, 145–152 (2014).
Leo, F. et al. Nature Photon. 4, 471–476 (2010).
Suh, M.-G., Yang, Q.-F., Yang, K. Y., Yi, X. & Vahala, K. J. Science 354, 600–603 (2016).
Puttnam, B. J. et al. Proc. Eur. Conf. Opt. Commun. http://doi.org/b7xx (2015).
Sacher, W. D. et al. Proc. Conf. Lasers Electro-Optics Paper JTh4C.3; http://doi.org/b7xz (2016).
Bao, C. et al. Opt. Lett. 39, 6126–6129 (2014).
Xue, X., Wang, P.-H., Xuan, Y., Qi, M. & Weiner, A. M. Laser Photon. Rev. 11, 1600276 (2017).
Temprana, E. et al. Science 348, 1445–1448 (2015).