Introduction

This work has focused on the development of a clock that uses a narrow transition (natural linewidth 375 Hz) in neutral calcium.  This transition, the ¹S₀ → ³P₁ intercombination line at 657 nm, is particularly well-suited for standards work, due to its convenient wavelength and extreme insensitivity to external fields and other perturbations.  In order to have ms interaction times and reduced Doppler shifts, we have constructed a compact magneto-optic trap for Ca atoms based on a frequency-doubled diode laser system at 423 nm [1].  With this trap we can load millions of atoms in tens of milliseconds, but in order to avoid unwanted Stark shifts, we actually release the atoms before probing the clock transition.  We perform the high resolution spectroscopy on the clock transition with a diode laser at 657 nm, which is locked tightly to a narrow fringe of an environmentally-isolated Fabry-Perot cavity (resultant laser linewidth ~ 1 Hz over 1 s).  To maintain a high signal-to-noise ratio at high resolution, we use Bordé-Ramsey saturation spectroscopy combined with a shelving detection technique.  With this approach we have resolved features as narrow as 200 Hz wide, although we usually work at resolutions of ~1 kHz for frequency standards work.  We then employ a feedback system to keep the laser frequency fixed on the atomic resonance.  Since optical frequencies are too high to count directly, we use a mode-locked fs laser comb to divide our calcium frequency down to the countable microwave regime.  This allows us to operate the system as a clock and to make direct comparisons with the NIST cesium fountain and other standards under development [2]. In fact, due to its comparatively simple and robust apparatus, the calcium standard has proven to be a useful flywheel for use by more accurate optical clocks (e.g., lattice clocks) to evaluate their systematics at the 10^-16 level [3,4].

Results and Prospects

With the laser-cooled calcium system we have demonstrated low instability (< 4x10^-15 @ 1s)[5] and an absolute inaccuracy of 3.3 Hz (at 456 THz) [6,7,8].  Our measured values for this transition are in good agreement with those measured by the calcium group at Physikalisch Technische Bundesanstalt (PTB).  To achieve the smallest uncertainty for the clock it was necessary to implement a second stage of cooling based on three-dimensional quenched narrow-line laser cooling that reduces the temperature of the atomic sample from 2 mK to below 10 μK [8,9]. However, since we now have a more accurate optical clock in the laboratory, we prefer to run the version of the clock based on a single stage of laser cooling. In this way we take advantage of the Ca clock's main strengths: high stability, short cycle time (2 ms), simplicity, quick set-up time (10 minutes) and robustness (> 12 hours run times without supervision). This version of the stabilized Ca clock is used in a variety of experiments including clock comparisons and generation of ultra-low noise microwaves.