Kramers-Kronig detection of four 20 Gbaud 16-QAM channels using Kerr combs for a shared phase estimation

Kaiheng Zou*, Peicheng Liao, Yinwen Cao, Arne Kordts, Ahmed Almaiman, Maxim Karpov, Martin Hubert Peter Pfeiffer, Fatemeh Alishahi, Ahmad Fallahpour, Moshe Tur, Tobias J. Kippenberg, Alan E. Willner

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


We experimentally demonstrate Kramers-Kronig detection of four 20 Gbaud 16-quadrature-amplitude-modulated (QAM) channels after 50 km fiber transmission using two soliton Kerr combs as signal sources and local oscillators. The estimated carrier phase at the receiver for each of the channels is relatively similar due to the coherence between the frequency comb lines. The standard deviation of the estimated carrier phase difference of the channels is less than 0.08 rad after 50 km single-mode fiber (SMF) transmission. This enables the carrier phase recovery derived from one channel to be shared among multiple channels. In the back-to-back scenario, the bit error rate (BER) performance for shared carrier phase recovery shows an optical signal-to-noise ratio penalty of ∼0.5 dB compared to the BER performance for carrier phase recovery when derived for each channel independently. BERs below the forward error correction threshold are achieved after 50 km SMF transmission with both independent and shared carrier phase recovery for four 20-Gbaud 16-QAM signals.

Original languageEnglish
Pages (from-to)1794-1797
Number of pages4
JournalOptics Letters
Issue number7
StatePublished - 1 Apr 2020


FundersFunder number
Basic Research Office of the Assistant Secretary of Defense for Research and Engineering
Centre for Micro-Nanotechnology
National Science FoundationY501119
Office of Naval ResearchN00014-16-1-2813
Air Force Office of Scientific ResearchFA9550-15-1-0166
Office of the Assistant Secretary for Research and Technology


    Dive into the research topics of 'Kramers-Kronig detection of four 20 Gbaud 16-QAM channels using Kerr combs for a shared phase estimation'. Together they form a unique fingerprint.

    Cite this