Where spin angular momentum goes during sub-picosecond demagnetization
In plain words
When a laser pulse demagnetizes nickel in under a trillionth of a second, the lost spin angular momentum must go somewhere, since angular momentum is conserved. Experiments show the crystal lattice eventually takes it, but how it gets there, and how much leaves as spin currents or electron orbital motion, is unclear.
Precise statement
In Ni, Co and Fe films pumped at absorbed fluences of order $1e4\,\mathrm{erg}/\mathrm{cm}^{2}$, magnetization drops within about 100 to 300 fs (Beaurepaire et al. 1996). Determine the time-resolved partition of the lost spin angular momentum among electron orbital moments, superdiffusive spin currents, magnons and the lattice (circularly polarized phonons, Einstein-de Haas rotation) during 0 to 500 fs, and the microscopic transfer channel (Elliott-Yafet spin-flip scattering, direct spin-phonon coupling). An answer is a quantitative angular-momentum budget versus time.
What would settle it
Simultaneous femtosecond measurement of spin and orbital moments (x-ray magnetic circular dichroism) and of phonon angular momentum (ultrafast diffraction) on one sample, closing the budget.
Status in the literature
Ultrafast diffraction showed the lattice receives the angular momentum on sub-picosecond times (Dornes et al. 2019) through circularly polarized phonons (Tauchert et al. 2022); the channel and shares in the first 100 fs remain open.