Light shift model errors in optical lattice clocks

Over recent decades, optical lattice clocks (OLCs) have made measurements with extreme precision. In large part, this precision is due to lattice confinement at the “magic wavelength,” where the AC Stark shift between the two clock states nearly cancels. Even so, residual lattice light shifts often limit the OLC's accuracy. An important detail to determining this shift is characterizing the atom's motional distribution in the lattice, which dictates the effective trap depth experienced by the average atom. Using experimental techniques like sideband thermometry or radial spectroscopy, a simplified lattice potential model (e.g., harmonic approximation) is employed to extract atomic temperature, which in turn parameterizes effective trap depth. Here, we explore potential model errors from this process, and present a novel experimental method to more directly determine the effective trap depth and residual light shifts. This work will help realize lattice light shift uncertainty <<10^-18.