Characterization of carrier-envelope offset phase noise generated by supercontinuum generation in microstructure fibers

Abstract

We present a new scheme for measuring distortions of the carrier-envelope offset (CEO) phase caused by supercontinuum generation in microstructure fibers. These distortions ultimately limit the accuracy of CEO phase stabilization schemes and absolute-phase dependent nonlinear optics experiments. The key advantages of the proposed scheme are the possibility to use a free running oscillator and the utilization of only one microstructure fiber. in the past 3 years. Virtually all measurement setups employ a self-referencing Measurement and control of the carrier-envelope offset rapidly progressed f-to-2f ~~ heterodyning scheme [I] and employ supercontinuum generation in a microstructure fiber. This set-up led to great progress in phase-sensitive nonlinear optics and precision. However, the strong nonlinearity of the fibers used for the supercontinuum generation process disturbs the CEO phase by means of AM-to-PM conversion. Thus, strictly speaking, only the pulses exiting the microstructure fiber are CEO stabilized. Depending on the magnitude of AM-to-PM conversion, the laser pulses exhibit increased CEO phase noise. To characterize the AM-to-PM conversion in to the fiber we compare twn differently measured CEO frequencies (Fig. I). One CEO frequency (subsequently referred to as 'green') is obtained by heterodyning green spectral components of the supercontinuum with its frequency doubled infrared part. This is the f-to-2f interferometer scheme commonly used for measuring the C.EO frequency. The second CEO frequency (subsequently referred to as 'blue') stems from heterodyning blue spectral componendaf the supercontinuum with frequency doubled oscillator pulses. As the oscillator pulses are not affected by possible AM-to-PM conversion in the microstructure fiber, comparing the two CEO frequencies reveals information about the CEO phase distortions introduced by the fiber. The CEO frequency of the free-running oscillator is coarsely adjusted to 30 MHz by tilting anintracavity glass slab. Both RF CEO frequencies, green and blue, are mixed with the same 29.8 MHz local oscillator and low-pass filtered, resulting in downshifted frequencies at about 200 kHz. These IF signals are directly digitized with 12 bit resolution at 6.7 MSds. The phase of both IF signals is retrieved by a Fourier filtering technique. Due to the symmetric mixing scheme phase shifts are conserved in the downshifted IF frequencies. In contrast to the measurement scheme utilized before [2], our setup does not require a stabilized CEO frequency and thus minimizes perturbations of the laser resulting from the stabilization loop. Moreover, using only one microstructure fiber avoids measurement errors resulting from