Reference-free, high-resolution measurement method of timing jitter spectra of optical frequency combs

www.nature.com/scientificreports OPEN received: 14 October 2016 accepted: 13 December 2016 Published: 19 January 2017 Reference-free, high-resolution measurement method of timing jitter spectra of optical frequency combs Dohyeon Kwon1, Chan-Gi Jeon1, Junho Shin1, Myoung-Sun Heo2, Sang Eon Park2,3, Youjian Song4 & Jungwon Kim1 Timing jitter is one of the most important properties of femtosecond mode-locked lasers and optical frequency combs. Accurate measurement of timing jitter power spectral density (PSD) is a critical prerequisite for optimizing overall noise performance and further advancing comb applications both in the time and frequency domains. Commonly used jitter measurement methods require a reference mode-locked laser with timing jitter similar to or lower than that of the laser-under-test, which is a demanding requirement for many laser laboratories, and/or have limited measurement resolution. Here we show a high-resolution and reference-source-free measurement method of timing jitter spectra of optical frequency combs using an optical fibre delay line and optical carrier interference. The demonstrated method works well for both mode-locked oscillators and supercontinua, with 2 × 10−9 fs2/Hz (equivalent to −174 dBc/Hz at 10-GHz carrier frequency) measurement noise floor. The demonstrated method can serve as a simple and powerful characterization tool for timing jitter PSDs of various comb sources including mode-locked oscillators, supercontinua and recently emerging Kerrfrequency combs; the jitter measurement results enabled by our method will provide new insights for understanding and optimizing timing noise in such comb sources. Optical frequency combs have evolved to be a powerful tool for various high-precision applications ranging from optical atomic clocks1 through frequency-domain spectroscopy2 to astro-combs3. Timing jitter (i.e., phase noise in pulse repetition-rate)4 is one of the most important properties of femtosecond mode-locked lasers and optical frequency combs. First, there are applications where the pulse timing jitter directly impacts the achievable performance, such as low-phase noise microwave generation5–7, timing synchronization for X-ray free-electron lasers8,9, pulse time-of-flight-based ranging10, photonic analogue-to-digital converters11, photonics-based radars12, and clock distribution networks13, to name a few. Timing jitter, in fact, also significantly contributes to the optical linewidth of optical frequency comb lines and phase noise of carrier-envelop-offset frequency (fceo) in the frequency domain14–16. Thus, accurate measurement of timing jitter power spectral density (PSD) is an important prerequisite for optimizing the jitter performance and further advancing frequency comb applications both in the time and frequency domains. However, complete characterization of timing jitter PSD in frequency combs (including mode-locked laser oscillators and their supercontinua) is often challenging and complicated. This difficulty frequently limited the accurate assessment of timing jitter performance of mode-locked lasers and frequency comb sources in laser laboratories. Conventional timing jitter measurement methods are based on microwave phase detector method17,18. Optical pulse train generated from the laser-under-test (LUT) is converted to a microwave signal using a high-speed photodiode and a band-pass filter, and its phase is compared with the reference microwave signals for phase noise (i.e., equivalent pulse timing jitter) measurement. This microwave phase detector method is very convenient, and even commercial signal source analysers can be used. However, measurement resolution is limited by both 1 School of Mechanical and Aerospace Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 34141, Korea. 2Center for Time and Frequency, Korea Research Institute of Standards and Science (KRISS), Daejeon 34113, Korea. 3Science of Measurements Program, University of Science and Technology (UST), Daejeon 34114, Korea. 4School of Precision Instruments and Opto-electronics Engineering, Tianjin University, Tianjin 300072, China. Correspondence and requests for materials should be addressed to J.K. (email: ) Scientific Reports | 7:40917 | DOI: 10.1038/srep40917 1 www.nature.com/scientificreports/ Figure 1. Schematic of the fibre delay-line-based timing jitter measurement method. PZT, piezo-electric transducer; FBG, fibre Bragg grating; EDFA, Erbium-doped fibre amplifier; FRM, Faraday rotating mirror; AOFS, acousto-optic frequency shifter; PD, photodetector; BPF, band-pass filter; PLL, phase-locked loop; DLL, delay-locked loop; HV amp, high voltage amplifier. the reference oscillator phase noise and additional phase noise added in the photo-detection process. Typical shot noise-limited measurement noise floor in photo-detection is ~−140 dBc/Hz level at 10-GHz carrier when detecting a ~100-MHz repetition-rate laser19. In addition, amplitude-to-phase conversion in the photodiode also adds excess timing jitter20–22. This noise floor is too high for timing jitter characterization of mode-locked lasers, because repetition-rate phase noise of free-running mode-locked lasers is often well below −140 dBc/Hz from 10 kHz Fourier frequency (when scaled to 10-GHz carrier)23–25. In order to improve the detection resolution, optical cross-correlation method19,23–25 can be used; this technique makes use of direct pulse-to-pulse timing comparison between two optical pulse trains using nonlinear optic processes (such as second-harmonic generation). When shorter pulse width and higher power are used, measurement noise floor of the optical cross-correlation can be improved. For example, when 50 mW average power and 60 fs pulse width were used, timing jitter could be measured with 3 ×  10−12 fs2/Hz (−202 dBc/Hz at 10-GHz carrier) background noise floor23. More recently developed optical heterodyne method, which measures timing jitter using optical spectrum interference between two identical mode-locked lasers, showed −212 dBc/Hz background noise floor at 10-GHz carrier26. Although the optical cross-correlation and optical heterodyne methods show the best timing jitter measurement resolution of mode-locked lasers, these methods require two identical mode-locked lasers or a reference mode-locked laser with matched repetition-rate and lower timing jitter, which is often a difficult requirement to meet. Thus, a high-resolution and reference-free timing jitter measurement method is highly desirable for simple yet accurate characterization of various optical frequency comb sources. In microwave engineering community, delay-line-based methods have enabled reference-source-free measurements of phase noise in microwave and radio-frequency (RF) oscillators27. This method can be directly applied to measure the timing jitter and phase noise of mode-locked lasers by using optical fibre link as the delay line. However, the measurement no (...truncated)