A Cavity-stabilized diode laser for dipole trapping of ytterbium

Abstract

Bad-cavity lasers using a gain medium with a narrower linewidth than the laser cavity have the potential to achieve very narrow linewidths and extremely long coherence times. Such lasers could serve as active frequency standards or enable very-long-baseline interferometric telescopes at optical frequencies. The 6s6p³P₀ to 6s²¹S₀ ground state transition in ¹⁷¹Yb is a promising candidate for the gain medium of a bad-cavity laser due to its 44 mHz linewidth. For ytterbium to be used efficiently as a gain medium, its inhomogeneous broadening must be suppressed to a level lower than the linewidth of its gain transition. In this thesis, I design, implement, and characterize an optical lattice trap for ytterbium atoms. The trap consists of a diode laser which is frequency stabilized to an adjustable-length cavity where the ytterbium atoms are trapped. The length of this cavity is then locked by comparison of the laser frequency to a stable reference cavity. The resulting standing wave has high enough intensity that the recoil energy of the gain transition is smaller than the energy spacing between motional modes of the trapped atoms. This situation is known as the Lamb-Dicke regime and means that there is an absence of recoil broadening. The large spacing between motional modes of the trap also enables sideband resolved cooling of the atoms, which allows cooling to temperatures of 3 [mu]K, near the ground state of the trapping potential. Additionally, if the wavelength of the optical lattice is chosen to be at the magic wavelength for ytterbium, where the relative AC Stark shift for the two levels of the gain transition is zero to first order, there is no broadening due to varying intensity in the trap. Since the Doppler effect, recoil broadening and the AC Stark shift are the main sources of inhomogeneous broadening, this trapping scheme is expected to suppress inhomogeneous broadening to a level of 1 Hz.