Optically trapped Feshbach molecules of fermionic $^{161}$Dy and $^{40}$K: Role of light-induced and collisional losses

TL;DR

This study dissects loss mechanisms in optically trapped DyK Feshbach molecules formed from fermionic Dy-161 and K-40, separating trap-light-induced decay from collisional losses across multiple near-infrared wavelengths. By spectroscopically probing four wavelength regions and tuning trap intensity, it identifies a low-loss window around ~2 μm where light-induced losses are minimized, while still enabling high densities. The authors observe Pauli suppression of inelastic dimer-dimer collisions near the Feshbach resonance, reducing β by roughly an order of magnitude, though magnetic-field inhomogeneities limit the full exploitation of this effect. The findings indicate that choosing appropriate trap wavelengths and improving magnetic-field control are key to achieving longer lifetimes and potential evaporative cooling of mass-imbalanced molecular samples, advancing studies of pairing and superfluidity in such systems.

Abstract

We study the decay of a dense, ultracold sample of weakly bound DyK dimers stored in an optical dipole trap. Our bosonic dimers are composed of the fermionic isotopes $^{161}$Dy and $^{40}$K, which is of particular interest for experiments related to pairing and superfluidity in fermionic systems with mass imbalance. We have realized dipole traps with near-infrared laser light in four different wavelength regions between 1050 and 2002 nm. We have identified trap-light-induced processes as the overall dominant source of losses, except for wavelengths around 2000 nm, where light-induced losses appeared to be much weaker. In a trap near 1550 nm, we found a plateau of minimal light-induced losses, and by carefully tuning the wavelength, we reached conditions where losses from inelastic collisions between the trapped dimers became observable. For very weakly bound dimers close to the center of a magnetically tuned Feshbach resonance, we demonstrate the Pauli suppression of collisional losses by about an order of magnitude.

Figures (5)

• Figure 1: Properties of the 7.3-G Feshbach resonance. (a) Scattering length $a$, (b) molecular state energy $E_{\rm mol}$, and (c) closed-channel fraction $Z$ as functions of the magnetic detuning $\delta B$. The insets in (b) and (c) focus on the near-universal region close to resonance.
• Figure 2: Trap-loss spectroscopy on DyK Feshbach molecules by variation of the ODT wavelength. Panel (a) refers to the region around 1051nm, with initial number $N_0=1.1e4$ and a hold time of 10ms. Panel (b) refers to the region around 1547nm, with $N_0=9e3$ and a hold time of 33ms. The error bars represent $1\sigma$ standard errors, calculated from multiple repetitions. The vertical dashed lines in (a) and (b) indicate the selected wavelengths 1051.03nm and 1547.12nm, respectively, used for all further measurements presented in this work.
• Figure 3: Loss rate as a function of the trapping beam peak intensity. We show the closed-channel one-body loss rates measured in the experiments for the 1051.03nm, 1547.12nm, and 2002.00nm ODTs in blue, red and gray dots, respectively. Note that the data and fit result for the $2002$-nm case have been scaled by a factor of 10 for better visibility. For comparison, we also show the data for a trap at 1064.04nm in yellow, already reported in Soave2023otf. The data are shown with 1$\sigma$ error bars derived from the lifetime fits (in some cases smaller than the symbol size). For conversion of intensity (in kW/cm$^2$) to optical potential depth (in µ K) use the factors 1.76, 1.81, 1.25, and 1.15 for the wavelength 1051 nm, 1064 nm, 1547 nm, and 2002 nm, respectively.
• Figure 4: Two-body loss rate coefficient versus magnetic detuning from resonance. The dashed line represents a one-parameter fit of Eq. \ref{['eq:FitPauli']} to the data points for detunings $\delta B<-20mG$ (filled symbols). Additional measurements (open symbols) closer to resonance, presumably beyond the limitations of our simple model, have not been taken into account for the fit. Vertical error bars indicate $1\sigma$ errors, derived from fits to the decay curves. The horizontal error bars represent the 2-mG peak-to-peak magnetic field noise.
• Figure 5: Collisional decay and the role of light-induced decay, observed in a 1547-nm trap. The red dots show the number of molecules measured after a variable hold time. The red dashed line represents a fit according to Eq. (\ref{['Eq:12bodyfit']}). The orange dot-dashed line shows a hypothetical loss curve with the same two-body loss in the complete absence of one-body losses.