Item

Abstract

A promising new method of attosecond spectroscopy, the attoclock, offers attosecond
resolution without requiring attosecond laser pulses. However, it requires knowledge
of the ionization time, opening up a long standing conceptual problem in physics (“how
much time does tunnelling take?”). In this work, the time delay in the tunnel ionization
process is considered. It is shown that a delay of the peak of the tunnelling wavefunction
exists as a matter of principle, and is caused by the interference of transmitted and
reflected components of the wavefunction. If sub-barrier reflections are neglected from
the wavefunction, the delay in the peak vanishes and tunnelling is seen to be instantaneous.
This is shown by considering a series of models of increasing complexity: the
square barrier, an adiabatically tunnelling electron, and a wavefunction based on the
Strong Field Approximation. This work has implications on the interpretation of attoclock
experiments: treatments based on instantaneous tunnelling should be adjusted in
order to achieve appropriate calibration of the attoclock.