Attosecond resolution from free running interferometric measurements

Quantitative uncertainty determination of phase retrieval in RABBITT

Reconstruction of attosecond beating by interference of two-photon transitions (RABBITT) is one of the most widely used approaches to measure the time delays in photoionization. The time delay, which corresponds to a phase difference of two oscillating signals, is usually retrieved by cosine fitting or fast Fourier transform (FFT). We propose two estimators for the phase uncertainty of cosine fitting from the signal per se of an individual experiment: (i) σ(φ _fit)≈B A2N, where B/A is the mean-value-to-amplitude ratio, and N is the total count number, and (ii) σ(φ _fit)≈1-R ² R ² n _bins, where n_bins is the total number of bins in the time domain, and R² is the coefficient of determination. The former estimator is applicable for the statistical fluctuation, while the latter includes the effects from various uncertainty sources, which is mathematically proven and numerically validated. This leads to an efficient and reliable approach to determining quantitative uncertainties in RABBITT experiments and evaluating the observed discrepancy among individual measurements, as demonstrated on the basis of experimental data.

Pump-Probe Delay Controlled by Laser-dressed Ionization with Isolated Attosecond Pulses

In a recent work [1], we demonstrated how laser-dressed ionization can be harnessed to control with attosecond accuracy the time delay between an extreme-ultraviolet (XUV) attosecond pulse train and an infrared (IR) femtosecond pulse. In this case, the comb-like photoelectron spectrum obtained by ionizing a gas target with the two superimposed beams exhibits peaks oscillating with the delay. Two of them can be found to oscillate in phase quadrature, allowing an optimal measurement and stabilization of the delay over a large range. Here we expand this technique to isolated attosecond pulses, by taking advantage of the delay-modulation of attosecond streaking traces. Although the photoelectron spectrum contains no peaks in that case, it is possible to reconstruct the pump-probe delay by simply monitoring the mean energy of the spectrum and the amplitude at this energy. In general, we find that active delay stabilization based on laser-dressed ionization is possible as long as the XUV pulses are chirped.