Although room-temperature CSACs based on the original NIST design have been commercialized and proven valuable for many uses, the devices have characteristics that can be significantly improved upon. For example, the cesium atoms in the vapor cells are accompanied by an inert buffer gas, which mitigates unwanted effects that occur when the atoms contact the glass cell wall. However, pressure shifts in the buffer gas result in output-signal frequency drift on the order of parts in 108 per year. That might seem to be a small variation, but much greater stability is needed, and NIST scientists believe that the devices can eventually achieve uncertainties three orders of magnitude smaller, or more.*
To reach that goal, NIST researchers are experimenting with using laser-cooled cesium** atoms in a high vacuum, which eliminates the need for buffer gas. It also increases the time during which the atoms can be interrogated because they are moving slower. The new design would also eliminate another source of error by using a pulsed laser to interrogate the atoms, as opposed to the continuous laser illumination of the original model.
But a high vacuum (< 10-7 Torr is the current target) customarily entails some seepage of contaminant gases through the cell’s glass enclosure. Many can be removed by the incorporation of “getter pumps” which sop up contaminants and maintain the vacuum. But getter pumps are less effective at trapping the common trace gases helium and methane.
The scientists plan to surmount this problem by using aluminosilicate glass, which is much less permeable by helium than, for example, the more familiar borosilicate glass (e.g., Pyrex). Ideally, this would give the instruments a lifetime of months to years.
* To understand the need for greater stability, consider that light travels at 0.3 m (about one foot) per nanosecond (10-9 sec). So, a light-based positioning system such as GPS that needs to be accurate to 3 m (10 feet) needs a light source that is accurate and stable to at least 10 ns.
** Cesium is, at present, used to realize time by international agreement. In the SI, the second is defined as “the duration of 9 192 631 770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium 133 atom.” NIST realizes that frequency for official U.S. time with an ultracold cesium fountain clock at an uncertainty around one part in 1016. The new cold-atom CSAC would have uncertainty in the range of parts in 1011.