Dual Frequency Comb Sampling of a Quasi-Thermal Incoherent Light Source F. R. Giorgetta, I. Coddington, E. Baumann, W. C. Swann, N. R. Newbury National Institute of Standards and Technology, 325 Broadway, Boulder, Colorado 80305 fabrizio.giorgetta@nist.gov Abstract: Dual, coherent frequency combs are used to measure the spectrum of an incoherent, quasi-thermal source through Fourier spectroscopy. The source spectrum is acquired over 1 THz bandwidth with an absolute frequency accuracy set by the combs. Work of the U.S. government, not subject to copyright. OCIS codes: 300.6300 (Spectroscopy, Fourier transforms), 320.7090 (Ultrafast lasers) One of the intriguing new applications of frequency combs is to the field of Fourier transform infrared spectroscopy (FTIR) [1-4]. To date, all comb-based spectroscopy systems have interrogated a passive sample, for example of a gas, where one comb probes the sample and the resulting complex spectrum is acquired by a second “local oscillator” (LO) comb that has a slightly different repetition rate. This system can be viewed as the analog of standard FTIR spectrometer, where the pulse train from one comb replaces the light in the fixed arm and the pulse train from the second comb replaces the light in the delayed arm of the Michelson interferometer. Here it is demonstrated that this dual-comb system can be reconfigured to measure the spectrum of an incoherent source in a manner analogous to standard passive FTIR systems. Just as with the previous spectroscopy demonstrations, this comb-based passive FTIR system has the advantages of a fast scanning rate, high frequency resolution, and a frequency accuracy that can reach kilohertz or even hertz levels depending on the underlying reference. However, due to its coherent single-mode detection and the photon statistics inherent to any thermal source, the signal-to-noise ratio of a single interferogram is limited to unity. Nevertheless, it is possible to generate spectra with reasonable signal-to-noise ratios through coherent signal averaging. This approach is interesting in terms of calibration of spectrometers or for precision measurements of frequency shifts in absorption lines across an incoherent spectrum. Fig. 1: Setup to measure the spectrum of an incoherent source with high frequency accuracy by sampling the spectrum in time with two asynchronous frequency combs. 16 nm BP: 16 nm wide bandpass, TBP: tunable BP, P: polarizer, PC: polarization controller, BD: balanced photodetector, LP: lowpass, M: number of pulses per frame, LO: local oscillator (frequency comb). Solid lines are fiber optic paths, dotted lines are electrical paths. The experimental setup is shown in Fig. 1. An amplified spontaneous emission (ASE) Er-fiber source serves as the quasi-thermal light source. Its output is spectrally filtered to 16 nm and passed through a cell of Hydrogen Cyanide (HCN) gas to add sharp spectral features. The dual-comb system is then used to measure the autocorrelation of the quasithermal ASE field, and therefore its spectrum, as follows. The pulse train of each LO comb is separately combined with the ASE light and the resulting overlap of the ASE electric field and the LO comb pulse is detected in a balanced detector. The detected voltage is proportional to the ASE field sampled by the comb pulses. Because the ASE light is incoherent, this voltage will be random. However, the phase-locks between the two LO sources are arranged such that the pulses from the two sources are optically coherent but have slightly different repetition rates, f r1 and f r2 . As a result, the pulse train from one source will slowly walk through the pulse train from the other source. The sampled voltages from the two pulse trains are then acquired as series of relative effective time delays; their product, generated by a mixer in Fig. 1, is exactly the autocorrelation of the ASE electric field (i.e., interferogram), modified by the spectral envelope of the comb pulses. This interferogram is generated repeatedly at a rate equal to the difference in comb repetition rates, ∆f r , at an effective time step ∆τ = f r1 -1 - f r2 -1 , over an effective time span of f r1 -1 . a605_1.pdf OSA / CLEO/QELS 2010 CTuS3.pdf