Influence of atomic recoil on the spectrum of resonance fluorescence from a two-level atom.

We have continued our precision studies of laser induced stimulated resonance Raman interactions in a sodium atomic beam with emphasis on Ramsey's method of separated oscillatory fields. We observed Raman-Ramsey fringes for a field separation of up to 30 cm, and the data were consistent with theoretical predictions. We have also been investigating the performance of a clock based on this interaction in a sodium atomic beam to determine the feasibility of such a scheme and to demonstrate any possible advantages over conventional microwave excited clocks. Recent performance showed a stability of 1 x 10-11 for a 5000 second averaging time. This compares favorably with commercial cesium clocks when difference in atom transit time and transition frequency are taken into consideration. Currently we are studying potential sources of long term frequency error in the Raman clock. Some of the error sources are similar to those in microwave clocks, such as the effects of path length phase shift, external magnetic fields, background slope, atomic beam misalignment and second order Doppler. The other error sources are unique to the Raman clock and include laser frequency detuning, laser intensity changes, laser beam misalignment, optical atomic recoil, the presence of nearby hyperfine levels, and other smaller effects. Our present investigations have centered around error sources that are unique to the Raman process. For example, we have found that laser detuning from resonance by 1% of the atomic linewidth can cause a fractional error of 2.4 x 10-11. Also a 1% change in laser intensity can generate an error of 2.5 x 10-12. Efforts are underway to find ways of reducing both detuning and intensity errors. In addition, laser misalignment can also be a source of significant error. A laser beam translation of 0.