Attosecond Angle-Resolved Photoelectron Spectroscopy

We report experiments on the characterization of a train of attosecond pulses obtained by high-harmonic generation, using mixed-color ($\mathrm{X}\mathrm{U}\mathrm{V}+\mathrm{I}\mathrm{R}$) atomic two-photon ionization and electron detection on a velocity map imaging detector. We demonstrate that the relative phase of the harmonics is encoded both in the photoelectron yield and the angular distribution as a function of XUV-IR time delay, thus making the technique suitable for the detection of single attosecond pulses. The timing of the attosecond pulse with respect to the field oscillation of the driving laser critically depends on the target gas used to generate the harmonics.

Figure 1

Overview of the experimental scheme. The output of a 50 Hz Ti:sapphire laser system (delivering near-transform limited 70 fs, 10 mJ laser pulses at 800 nm) is split into an annular outer beam (adjustable up to a diameter ${\varnothing }_{\mathrm{o}\mathrm{u}\mathrm{t}}=20\mathrm{m}\mathrm{m}$) and an inner beamlet ($\varnothing =3\mathrm{m}\mathrm{m}$), which are delayed with respect to each other by passage through two pieces of glass cut from the same antireflection-coated plate. An $f=2\mathrm{m}$ lens focuses the copropagating IR beams into a rare gas jet, where the annular beam generates harmonics. A masking plate with a $\varnothing =0.5\mathrm{m}\mathrm{m}$ pinhole blocks the annular part of the IR beam 20 cm behind the focus, and transmits the inner IR beam and the harmonics into a spectrometer chamber. There the lasers cross an atomic beam and the resulting photoelectrons are detected using velocity imaging, which projects the electrons with the aid of electrostatic fields onto a position-sensitive channel-plate detector. The 3D velocity distribution is recovered from the measured projection by applying an iterative procedure [10].

Figure 2

(top panels) Slice through the 3D velocity distribution measured for mixed-color ($\mathrm{X}\mathrm{U}\mathrm{V}+\mathrm{I}\mathrm{R}$) two-photon ionization of argon, where the XUV light has been obtained by high-harmonic generation in argon. The 3D velocity distribution contains a series of prominent rings resulting from ionization by one XUV photon (11th–21st harmonic) and weaker sidebands at intermediate energies where two mixed-color ionization processes involving the adjacent harmonics interfere. Arrows in the measurement on the right indicate photoelectron signals due to one-photon ionization with the 13th harmonic ($H13$) and the sideband resulting from two-photon ionization involving the 13th and the 15th harmonic ($S14$). In the image on the left, the interference is destructive for most sidebands, whereas in the image on the right it is constructive. (bottom panel) Extracted parameters $\alpha$, $\beta$, and $\gamma$ [defined in Eq. (1)] for sideband $S16$ that contains interfering contributions from ionization by the 15th and 17th harmonic, as a function of the time delay between the XUV pulse and the IR dressing beam. Both parameter $\alpha$, describing the intensity of the sideband, and parameters $\beta$ and $\gamma$, describing the angular distribution of the ejected photoelectrons in the sideband, oscillate as a function of the delay with half the IR period, and allow the extraction of a phase that contains information about the phase shift between the 15th and 17th harmonic. The total signal oscillates with the full IR period.

Figure 3

(top panel) Extracted phase differences ${\phi }_{\alpha }(q+1)-{\phi }_{\alpha }(q-1)$ between adjacent harmonics for three experiments performed using argon as a harmonic generation gas (open triangles), and using xenon (open circles) and krypton (open squares) as harmonic generation gases. The three argon data sets were taken on different days under conditions as similar as we could make them. (bottom panel) XUV temporal shape for argon, krypton, and xenon as harmonic generation gases, on the basis of the harmonic phase differences shown in the top plot. The harmonic amplitudes needed for the reconstruction were obtained from photoelectron spectra recorded in the absence of the IR dressing beam.

Figure 4

Difference between the extracted phases ${\varphi }_{\alpha }$ and ${\varphi }_{\beta }$ for three experiments using argon (solid squares), krypton (solid circles), and xenon (open squares) as harmonic generation gas, along with a theoretical prediction based on the mixed-color two-photon ionization model (open circles). Since the phase difference depends only weakly on the harmonic amplitudes and is independent of the harmonic phases, the experimental results lie on top of each other, and agree reasonably well with the theoretical prediction.