Figure 1. Broad applications of optical microcombs. LiDAR, light detection and ranging; IM-DD, intensity modulation–direct detection.

Figure 2. (a) Chronology for different applications of optical microcombs, including the first demonstrations of microwave frequency synthesis [20], microwave photonic filtering [29], optical communications [41], microwave photonic signal processing [34], generation of multiphoton quantum states [59], dual-comb spectroscopy [60], ranging [46,47], single neuron [54], astrocombs [50,51], and convolutional neural networks [55,56]. (b) Number of publications on different applications of optical microcombs in Science Citation Index journals versus year since 2012. Data were taken from ISI Web of Science.

Figure 3. Development roadmap of optical microcombs based on various material platforms. Values of wavelength ranges are taken from: Ref. [10] for 2007, Ref. [96] for 2008, Ref. [75] for 2009, Ref. [68] for 2010 (Hydex), Ref. [69] for 2010 (Si₃N₄), Ref. [63] for 2011 (SiO₂), Ref. [72] for 2011 (MgF₂), Refs. [20,97] for 2012, Ref. [73] for 2013 (MgF₂), Ref. [71] for 2013 (SiO₂), Ref. [98] for 2014 (AlN), Ref. [85] for 2014 (diamond), Ref. [65] for 2014 (MgF₂), Ref. [79] for 2015 (Si), Ref. [74] for 2015 (MgF₂), Ref. [99] for 2015 (Si₃N₄), Ref. [86] for 2016 (AlGaAs), Ref. [82] for 2016 (Si₃N₄), Ref. [80] for 2016 (Si), Refs. [100,101] for 2017 (Si₃N₄), Ref. [102] for 2017 (SiO₂), Refs. [103,104] for 2018 (Si₃N₄), Ref. [105] for 2018 (Hydex), Refs. [94,95] for 2019 (LiNbO₃), Ref. [106] for 2019 (Hydex), Ref. [107] for 2019 (MgF₂), Ref. [108] for 2020 (LiNbO₃), Ref. [109] for 2020 (SiO₂), Refs. [110,111] for 2020 (Si₃N₄), Ref. [88] for 2020 (AlGaAs), Ref. [92] for 2020 (GaP), Ref. [112] for 2021 (Hydex), Ref. [90] for 2021 (SiC), Ref. [113] for 2021 (AlN), Ref. [114] for 2021 (Si₃N₄), and Ref. [91] for 2021 (Ta₂O₅).

Figure 4. Recent progress in optical microcomb generation based on new material platforms. (a) Aluminum gallium arsenide (AlGaAs). Reprinted from [87] under a CC-BY license. (b) Deuterated silicon nitride (D-SiN). Reprinted with permission from [89] © The Optical Society. (c) Silicon carbide (SiC). Reprinted from [90] under a CC-BY license. (d) Tantalum pentoxide (Ta₂O₅). Reprinted with permission from [91] © The Optical Society. (e) Gallium phosphide (GaP). Reprinted by permission from Nature Publishing Group: Wilson et al., Nat. Photonics 14, 57, (2020) [92]. (f) Lithium niobate (LiNbO₃). Reprinted by permission from Nature Publishing Group: Zhang et al., Nature 568, 373 (2019) [93]. (g) Silicon nitride (Si₃N₄) with gated graphene. Reprinted by permission from Nature Publishing Group: Yao et al., Nature 558, 410 (2018) [103]. In (a)‒(g), (i) shows the microresonators used for generating microcombs and (ii) shows the generated comb spectra.

Figure 5. Recent progress in high-volume manufacturing of optical microcomb devices. (a) Microcomb generator based on integrated silica ridge resonators with ultrahigh Q factors. Reprinted by permission from Nature Publishing Group: Yang et al., Nat. Photonics 12, 297 (2018) [138]. (b) Battery-operated microcomb generator. Reprinted by permission from Nature Publishing Group: Stern et al., Nature 562, 401 (2018) [139]. (c) Electrically pumped microcomb generation by using a commercial InP laser diode. Reprinted from [140] under a CC-BY license. (d) Heterogeneously integrated microcomb generators on a silicon substrate. Reprinted from [135] under a CC-BY license. (e) A microcomb generator based on a high-power, low-noise laser fully integrated with a Si₃N₄ MRR. Reprinted from [136] under a CC-BY license.

Figure 6. Recent progress in new device design for generating optical microcombs. (a) Waveguide geometry engineering in a Si₃N₄ MRR. Reprinted from [104] under a CC-BY license. (b) Dispersive waves engineering in a Si₃N₄ MRR. Reprinted with permission from [100] © The Optical Society. (c) Dispersion oscillation engineering in a tapered Si₃N₄ MRR with varied waveguide width. Reprinted from [152] under a CC-BY license. (d) Coupling engineering based on a deformed SiO₂ microtoroid resonator. Reprinted from [109] under a CC-BY license. (e) Multi-mode interaction engineering in two coupled Si₃N₄ MRRs. Reprinted by permission from Nature Publishing Group: Helgason et al., Nat. Photonics 15, 305 (2021) [151]. (f) Multi-mode interaction engineering in a Si₃N₄ concentric MRR. Reprinted from [150] under a CC-BY license.

Figure 7. Recent progress in microcomb new driving mechanisms. (a) Access range extension of soliton microcombs generated by a silica microrod resonator by using an auxiliary laser. Reprinted from [145] under a CC-BY license. (b) Spectral bandwidth broadening of soliton microcombs generated by a silica microtoroid resonator via bichromatic pumping. Reprinted from [141] under a CC-BY license. (c) Piezoelectric-controlled DKS generation in a Si₃N₄ MRR. Reprinted by permission from Nature Publishing Group: Liu et al., Nature 583, 385 (2020) [198]. (d) Turnkey soliton microcomb generation in a Si₃N₄ MRR. Reprinted from [110] under a CC-BY license. (e) Self-emergence robust soliton microcomb generation in a Hydex MRR. Reprinted from [192] under a CC-BY license. (f) DKS generation in both a Si₃N₄ MRR and a MgF₂ crystalline resonator pumped by optical pulses generated from a GSL. Reprinted from [147] under a CC-BY license. (g) Soliton-based resonant supercontinuum generation in a Si₃N₄ microresonator pumped by optical pulses. Reprinted with permission from [199] © The Optical Society. (h) Synchronization of two soliton microcombs in two Si₃N₄ MRRs. Reprinted by permission from Nature Publishing Group: Jang et al., Nat. Photonics 12, 688 (2018) [197].

Figure 8. (a) Allan deviations (at an integration time of 1 s) and (b) SSB phase noise (at 1-kHz, 10-kHz, and 10-MHz offset frequencies) versus synthesized frequencies for frequency synthesizers based on different microcombs as well as LFCs generated by solid-state lasers (SSLs) and mode-locked fiber lasers (MLFLs). Values of Allan deviations and SSB phase noise are taken from Ref. [24] for the 16.82-GHz, 6.73-GHz, and 7.05-GHz signals synthesized based on microcombs generated by a Si₃N₄ MRR, Ref. [27] for the 100-GHz signal synthesized by microcombs generated by a Si₃N₄ MRR, and Ref. [28] for the 300-GHz signal synthesized by microcombs generated by a Si₃N₄ MRR. Values of Allan deviations are taken from Ref. [24] for the 4.7-GHz signal synthesized based on microcombs generated by a MgF₂ whispering-gallery-mode (WGM) resonator, Ref. [16] for the 0.01-GHz signal synthesized based on LFCs generated by a solid-state laser, Ref. [206] for the 10-GHz signal synthesized based on LFCs generated by two SSLs, Ref. [207] for the 10-GHz signal synthesized based on LFCs generated by two MLFLs, Ref. [26] for the 10-GHz and 20-GHz signals synthesized based on microcombs generated by two Si₃N₄ MRRs, Ref. [22] for the 14-GHz signal synthesized based on microcombs generated by a MgF₂ WGM resonator, Ref. [144] for the 22-GHz signal synthesized by microcombs generated by a SiO₂ wedge resonator, values are taken from Thorlabs, Inc. for the 0.1-GHz [208] and the 0.25-GHz [209] signals synthesized by commercial LFCs generated by mode-locked fiber lasers. Values of SSB phase noise are taken from Ref. [210] for the 11.55-GHz signal synthesized based on LFCs generated by two MLFLs, Ref. [23] for the 14-GHz signal synthesized based on microcombs generated by a MgF₂ WGM resonator, Ref. [211] for 18-GHz, 24-GHz, 30-GHz, and 36-GHz signal synthesized based on microcombs generated by a MgF₂ WGM resonator, Ref. [25] for 20-GHz signal synthesized based on microcombs generated by a Hydex MRR, Ref. [20] for 22-GHz signal synthesized based on microcombs generated by a SiO₂ microdisk resonator.

Figure 9. Recent progress in microcomb-based frequency synthesizers. (a) Microwave frequency synthesizers in the X and K bands based on soliton microcombs generated by Si₃N₄ MRRs. Reprinted by permission from Nature Publishing Group: Liu et al., Nat. Photonics 14, 486 (2020) [26]. (b) 100-GHz microwave frequency synthesizer based on soliton microcombs generated by a Si₃N₄ MRR [27]. (c) Microwave frequency synthesizer using a transfer oscillator approach to divide the repetition rate of a soliton microcomb [23]. (d) Microwave frequency synthesizer realized by injecting a soliton microcomb into a GSL [24]. (b), (c), and (d) Reprinted under a CC-BY license. (e) Optical frequency synthesizer with f–2f stabilization based on two soliton microcombs generated by a Si₃N₄ MRR and a SiO₂ microdisk. Reprinted by permission from Nature Publishing Group: Spencer et al., Nature 557, 81 (2018) [21].

Figure 10. Influence of tap number on the performance of microcomb-based microwave photonic filters. (a) Low-pass filters with the same cut-off frequency of 1 GHz but different tap numbers ranging from 10 to 80. (b) Band-pass filters with the same center frequency of 10 GHz but different tap numbers ranging from 10 to 80. In (a) and (b), (i), (ii), and (iii) show the simulated transmission spectra, the filtering resolutions (i.e., 3-dB bandwidth) versus tap numbers, and the MSSRs versus tap numbers, respectively.

Figure 11. Recent progress in microcomb-based microwave photonic filters. (a) 80-tap microwave photonic transversal filter based on soliton crystal microcombs generated by a Hydex MRR. (i) and (ii) show the measured spectral responses of bandpass filters with a tunable center frequency and a tunable bandwidth, respectively. © 2019 IEEE. Reprinted, with permission, from Xu et al., J. Lightwave Technol. 37, 1288 (2019) [30]. (b) Microwave photonic transversal filter based on bandwidth scaling. (i) and (ii) show the measured spectral responses of a bandpass filter with various tap numbers and a bandpass filter with tunable center frequency, respectively. Reprinted from [31] under a CC-BY license. (c) Reconfigurable microwave photonic filters without external pulse shaping based on rich soliton states formed in a Si₃N₄ MRR. Reprinted from [32] under a CC-BY license. (d) A reconfigurable microwave photonic filter using soliton microcombs generated by an AlGaAs-on-insulator source chip to drive a silicon-in-insulator (SOI) shaping chip. Reprinted from [33] under a CC-BY license.

Figure 12. Influence of tap number on the performance of microcomb-based microwave photonic signal processors. (a) First-order differentiators with different tap numbers ranging from 10 to 80. (b) 0.5-order differentiators with different tap numbers ranging from 10 to 80. In (a) and (b), (i), (ii), (iii), and (iv) show the simulated amplitude responses, phase responses, the output waveforms for Gaussian input signal, and the root-mean-square errors (RMSEs) between the output waveforms and the ideal results versus tap numbers, respectively.

Figure 13. Recent progress in realizing new basic computing functions using microcomb-based microwave photonic signal processors. (a) Theoretical and measured output temporal waveforms of a fractional-order differentiator. © 2020 IEEE. Reprinted, with permission, from Tan et al., IEEE Trans. Circuits Syst. II-Express Briefs 67, 2767 (2020) [37]. (b) Theoretical and measured output temporal waveforms of an integrator. © 2020 IEEE. Reprinted, with permission, from Xu et al., IEEE Trans. Circuits Syst. II-Express Briefs 67, 3582 (2020) [38]. (c) Theoretical and measured output temporal waveforms of a fractional-order Hilbert transformer. © 2019 IEEE. Reprinted, with permission, from Tan et al., J. Lightwave Technol. 37, 6097 (2019) [35]. (d) Theoretical and measured spectral response of a fractional-order Hilbert transformer with tunable (i) bandwidth and (ii) center frequency. © 2021 IEEE. Reprinted, with permission, from Tan et al., J. Lightwave Technol. 39, 7581 (2021) [36].

Figure 14. Recent progress in realizing other functions using microcomb-based microwave photonic signal processors. (a) Microwave photonic phase encoding. © 2020 IEEE. Reprinted, with permission, from Xu et al., J. Lightwave Technol. 38, 1722 (2020) [40]. (b) AWG: theoretical and measured tunable RF square and sawtooth waveforms. © 2020 IEEE. Reprinted, with permission, from Tan et al., J. Lightwave Technol. 38, 6221 (2020) [39]. (c) AWG: theoretical and measured triangle and “UVA”-like waveforms. Reprinted with permission from [225]. © The Optical Society.

Figure 15. Coherent optical communications based on optical microcombs. (a) ∼55.0 Tbit/s coherent communication over 75 km based on two interleaved soliton microcombs generated by Si₃N₄ MRRs. Reprinted by permission from Nature Publishing Group: Marin-Palomo et al., Nature 546, 274 (2017) [42]. (b) ∼4.4 Tbit/s coherent communication over 80 km based on a mode-locked dark-pulse microcomb generated by a normal-dispersion Si₃N₄ MRR. Reprinted from [43] under a CC-BY license. (c) ∼44.2 Tbit/s coherent communication over 76.6 km based on a soliton crystal microcomb generated by a Hydex MRR. Reprinted from [44] under a CC-BY license. (d) ∼1.68 Tbit/s coherent communication over 50 km based on coherence-cloned soliton microcombs generated by Si₃N₄ MRRs. Reprinted from [252] under a CC-BY license.

Figure 16. IM-DD optical communications based on optical microcombs. (a) ∼1.45 Tbit/s data transmission over 40 km based on a single soliton microcomb generated by a MgF₂ crystalline microresonator. Reprinted with permission from [45]. © The Optical Society. (b) ∼2 Tbit/s PAM4 data transmission over 2 km based on an AlGaAs-on-insulator comb source and an SOI transmitting–receiving chip. Reprinted from [33] under a CC-BY license.

Figure 17. DCS based on optical microcombs. (a) DCS measurement of dichloromethane’s absorption by using two microcombs on a single chip with a single pump. Reprinted from [260] under a CC-BY license. (b) DCS measurement of acetone in the liquid phase based on mid-infrared microcombs generated by two silicon MRRs. Reprinted from [148] under a CC-BY license. (c) DCS measurement of methane gas’s absorption based on mid-infrared combs generated by combining a 1.5-μm soliton microcomb and a 1.0-μm EO frequency comb to pump a periodically poled lithium niobate (PPLN) crystal. Reprinted with permission from [261] © The Optical Society. (d) DCS measurement of a mixture of methane and ethane gas’s absorption based on mid-infrared combs generated by 1.55-μm counter-propagating soliton microcombs. Reprinted from [265] under a CC-BY license. (e) Gas molecule detection based on three Stokes solitons generated by a graphene-deposited silica microsphere resonator. Reprinted from [274]under a CC-BY license.

Figure 18. Ranging based on optical microcombs. (a) Ultrafast ranging of moving targets based on soliton microcombs generated by two Si₃N₄ MRRs. (b) Distance measurement based on counterpropagating soliton microcombs generated by a single silica wedge resonator. (c) Spectrally resolved laser ranging with nanometric precision based on a soliton microcomb generated by a tapered Si₃N₄ MRR. (d) Massively parallel ranging using a soliton microcomb generated by a Si₃N₄ MRR pumped by chirped laser. (a) From [46]. Reprinted with permission from AAAS. Trocha et al., Science 359, 887–891 (2018). (b) From [47]. Reprinted with permission from AAAS. Suh and Vahala, Science 359, 884–887 (2018). (c) Figure 1 reprinted with permission from Bao et al., Phys. Rev. Lett. 126, 023903 (2021) [266]. Copyright 2021 by the American Physical Society. (d) Reprinted by permission from Nature Publishing Group: Riemensberger et al., Nature 581, 164 (2020) 48]. Copyright 2020.

Figure 19. Astrocombs based on optical microcombs. (a) A spectrograph calibrator based on a soliton microcomb generated by a SiO₂ microdisk resonator. (b) A microphotonic astrocomb based on soliton microcomb generated by a Si₃N₄ MRR. (a) Reprinted by permission from Nature Publishing Group: Suh et al., Nat. Photonics 13, 25–30 (2019) [51]. Copyright 2019. (b) Reprinted with permission from Nature Publishing Group: Obrzud et al., Nat. Photonics 13, 31–35 (2019) [50]. Copyright 2019.

Figure 20. Frequency measurements based on optical microcombs. (a) Detecting repetition rates of microcombs based on Vernier frequency division. (b) Rapid and broadband optical frequency measurement based on a Vernier spectrometer with two counterpropagating soliton microcombs generated by a single silica wedge resonator. (c) Measuring the RF spectra of microcombs from the microwave to terahertz bands based on an all-optical RF spectrum analyzer. (d) Measuring the relative frequency drift between two optical cavities based on a soliton microcomb generated by a silica microdisk resonator and subsequently broadened via coherent supercontinuum generation after passing a Si₃N₄ waveguide. (a) Reprinted with permission from [268] under a CC-BY license. (b) From Yang et al., Science 363, 965–968 (2019) [143]. Reprinted with permission from AAAS. (c) Reprinted with permission from [269]. © Optica Publishing Group. (d) Figure 4 reprinted with permission from Lamb et al., Phys. Rev. Appl. 9, 024030 (2018) [267]. Copyright 2018 by the American Physical Society.

Figure 21. Microwave spectrum channelizers based on optical microcombs. (a) Microwave spectrum channelizer with 20 channels in the C band based on microcombs generated by a 200-GHz FSR Hydex MRR. (b) Microwave spectrum channelizer with 92 channels in the C band based on soliton crystal microcombs generated by a 49-GHz FSR Hydex MRR. (a) © 2018 IEEE. Reprinted, with permission, from Xu et al., J. Lightwave Technol. 36, 4519 (2018) [270]. (b) © 2020 IEEE. Reprinted, with permission, from Xu et al., J. Lightwave Technol. 38, 5116 (2020) [271].

Figure 22. Reconfigurable single neuron based on optical microcombs. (a) Experimental setup. (b) Experimental results for classification of handwritten digits and cancer cells. Reprinted with permission from Feldmann et al., Laser Photon. Rev. 14, 2000070 (2020) [54].

Figure 23. CNNs based on optical microcombs. (a) Schematic of a convolution computation accelerator. Parallel operations are achieved by using multiple wavelengths derived from a soliton microcomb generated by a Si₃N₄ MRR. (b) Operation principle of an optical CNN based on time–wavelength interleaving using a soliton crystal microcomb generated by a Hydex MRR. (a) Reprinted with permission from Nature Publishing Group: Feldmann et al., Nature 589, 52–58 (2021) [55]. Copyright 2021. (b) Reprinted with permission from Nature Publishing Group: Xu et al, Nature 589, 44–51 (2021) [56]. Copyright 2021.

Figure 24. Generation of single/entangled photons based on optical microcombs. (a) Generation of heralded single photons based on a Hydex MRR embedded in an external fiber cavity. (b) Generation of orthogonally polarized photon pairs based on SFWM in a bi-chromatically pumped Hydex MRR. (c) Generation of bi- and multiphoton-entangled qubits based on a microcomb generated by a Hydex MRR. (d) Generation of energy–time entangled states based on a biphoton microcomb generated by a Si₃N₄ MRR. (e) Generation of high-dimensional entangled quantum states based on SFWM in a Hydex MRR. (f) Generation of qubit and qutrit frequency-bin entanglement based on a microcomb generated by a Si₃N₄ MRR. (a) Reprinted with permission from [317]. © Optica Publishing Group. (b) Reprinted from [318] under a CC-BY license. (c) From Reimer et al., Science 351, 1176–1180 (2016) [59]. Reprinted with permission from AAAS. (d) Reprinted with permission from [319].© Optical Society of America. (e) Reprinted with permission from Nature Publishing Group: Kues et al., Nature 546, 622-626 (2017) [320]. Copyright 2017. (f) Reprinted with permission from [321]. © Optical Society of America.

Figure 25. Generation of squeezed light based on optical microcombs. (a) All-optical squeezing based on a Si₃N₄ MRR pumped above the power threshold of optical parametric oscillation. (b) Generation of squeezed vacuum states based on parametric down-conversion in a LiNbO₃ WGM resonator. (c) Generation of (i) quadrature squeezing and (ii) photon number difference squeezing based on SFWM in a Si₃N₄ MRR. (d) Generation of strongly squeezed light based on SFWM in a photonic molecule resonator consisting of two coupled Si₃N₄ MRRs. (e) Generation of squeezed quantum microcomb in a silica wedge resonator. (a) Figures 1 and 2 reprinted with permission from Dutt et al., Phys. Rev. Appl. 3, 044005 (2015) [326]. Copyright 2015 by the American Physical Society. (b) Reprinted with permission from [327]. © Optical Society of America. (c) Reprinted from [328] under a CC-BY license. (d) Reprinted from [329] under a CC-BY license. (e) Reprinted from [330]under a CC-BY license.

Figure 26. Comparison of the state-of-the-art material platforms for optical microcomb generation. (a) Materials’ bandgaps and cut-off wavelengths versus their Kerr nonlinear coefficients (n₂). (b) Microresonators’ quality (Q) factors versus the materials’ n₂. (c) Materials’ bandgaps and cut-off wavelengths versus microresonators’ FSRs in the microwave frequency band (0.3‒300 GHz). Values of n₂, bandgap and cut-off wavelength are taken from Ref. [83] for MgF₂ and CaF₂, Ref. [91] for Ta₂O₅, Ref. [92] for SiO₂, diamond, Hydex, AlN, Si₃N₄, SiC, Si, GaP, and AlGaAs, and Ref. [95] for LiNbO₃. Values of Q factors are taken from Refs. [65,73,74] for MgF₂, Refs. [21,63,102,109,331] for SiO₂, Refs. [75,76] for CaF₂, Ref. [85] for diamond,Refs. [93,95,108] for LiNbO₃, Refs. [68,112] for Hydex, Ref. [98] for AlN, Refs. [81,82,99,114,139] for Si₃N₄, Ref. [91] for Ta₂O₅, Ref. [90] for SiC, Refs. [79,332] for Si, Ref. [92] for GaP, and Refs. [86–88] for AlGaAs. Values of FSRs are taken from Refs. [65,73,74] for MgF₂, Refs. [21,63,102] for SiO₂, Refs. [75,76] for CaF₂, Ref. [85] for diamond, Refs. [93,95,108] for LiNbO₃, Refs. [30,68,333] for Hydex, Ref. [98] for AlN, Refs. [26,69,81,82,99,114,139] for Si₃N₄, Ref. [90] for SiC, Refs. [79] for Si, Ref. [92] for GaP, and Ref. [87] for AlGaAs.

Figure 27. (a) Schematic of the state-of-the-art microcomb-based microwave photonic transversal filter system. (b) Schematic showing the concept of a monolithically integrated microcomb-based microwave photonic transversal filter system. CW laser, continuous-wave laser; PC, polarization controller; EDFA, erbium-doped fiber amplifier; MRR, microring resonator; TCS, temperature control stage; OSS, optical spectral shaper; EOM, electro-optic Mach–Zehnder modulator; SMF, single-mode fiber; OC, optical coupler; PD, photodetector; OSA, optical spectrum analyzer; ART, anti-reflection termination.