Interferometric measurements shake up attosecond science

When an atom absorbs light and ejects an electron, the emission doesn’t happen instantaneously. Moreover, the photoionization delay time depends on the atomic orbital from which the electron was liberated. That was the conclusion of a landmark 2010 experiment at the Max Planck Institute of Quantum Optics in Garching, Germany. The researchers there ionized neon atoms with subfemtosecond extreme UV (XUV) pulses; electrons freed from 2p orbitals were distinguishable from those ejected from 2s orbitals by their 30 eV difference in kinetic energy. The experiment probed only the difference between the two photoionization delays, not the absolute value of either one. But the nonzero value of that difference had important implications for the emerging field of attosecond science: The time at which a photoelectron is released isn’t a reliable measure of the time an attosecond pulse arrives at its target.

But of the 21 ± 5 attosecond difference in delay times measured in the Garching experiment, theory could account for just 9 attoseconds—a value that’s been reconfirmed as theorists have refined their calculations. Now Marcus Isinger, his thesis adviser Anne L’Huillier , and their collaborators at three Swedish universities have resolved the discrepancy. The answer lies in so-called shake-up, the simultaneous removal of one electron and excitation of another. In particular, the shake-up process in Ne that removes one 2p electron and promotes another to a 3p orbital differs in energy from direct 2s ionization by just 7.4 eV—less than the Garching experiment’s energy resolution.

To study the processes separately, Isinger and company used an interferometric technique developed in 2001. By superposing a series of phase-locked high harmonics, they produced a train of attosecond XUV pulses, represented in green in the figure, and combined it with an IR wave (purple) whose frequency was half the interharmonic spacing. Sideband frequencies midway between successive harmonics can be produced in two ways—combining XUV absorption with either absorption or emission of an IR photon—and the intensities of sideband-initiated processes oscillate when the phase between the XUV and IR waves is varied. Because of the spectral resolution of the frequency comb, direct 2s ionization and the shake-up process, though close in energy, produced well-resolved peaks with very different oscillation amplitude and phase. With the shake-up signal thus isolated, the measured time-delay difference between 2s and 2p ionization agrees well with theory. The results put attosecond time-delay techniques on firmer footing to address larger atoms and more complicated systems. (M. Isinger et al., Science 358, 893, 2017 .)