Alkali pressure shifts and their temperature dependence: Measurements with the Rb isoclinic point

Accurately assessing temperature variations of vapor-phase systems is a non-trivial problem, and one that has significant implications for the long-term stability of atomic frequency standards. As an example, consider the straightforward case of attempting to measure and stabilize the vapor temperature of Rb in a gas-cell atomic clock. Routinely, the temperature of the vapor is measured by placing a thermocouple or thermistor at the cold point of the resonance cell´s exterior (i.e., the region of the liquid Rb pool). As a consequence of vapor-phase/condensed-phase equilibrium, this temperature defines the Rb vapor density and by implication its mean kinetic energy. However, using a temperature probe in this manner has at least two problems: 1) the temperature of the vapor is not measured directly, only the temperature at a point on the exterior of the vapor´s glass container is measured, and 2) the temperature gradients that likely exist over the vapor´s volume are ignored. One can, of course, place additional probes on the cell´s exterior, but this still leaves the question of how cell-exterior temperature gradients map to interior-vapor gradients. Therefore, in order to improve the long-term stability of next-generation atomic clocks, there is a need for more accurate means of assessing actual vapor temperatures.