Zero-Dead-Time Operation of Interleaved Atomic Clocks

We demonstrate a zero-dead-time operation of atomic clocks. This clock reduces sensitivity to local oscillator noise, integrating as nearly $1/\tau$ whereas a clock with dead time integrates as $1/{\tau }^{1/2}$ under identical conditions. We contend that a similar scheme may be applied to improve the stability of optical clocks.

Figure 1

Ramsey $\pi /2$ pulse timing diagram of a ZDT implementation of two clocks (${\mathrm{SC}}_1$ and ${\mathrm{SC}}_2$). In each clock cycle the first (second) $\pi /2$ pulse of one clock coincides with the second (first) $\pi /2$ pulse of the other clock. Typical clock processes of detection and state preparation occur during the times between LO interrogation.

Figure 2

High level schematic of our ZDT demonstration. Two identical atomic clocks, ${\mathrm{SC}}_1$ and ${\mathrm{SC}}_2$, alternately monitor the LO in the presence of an added phase noise. The combined outputs continuously track the resulting phase evolution.

Figure 3

Typical Ramsey clock fringe indicating the system noise floor. With $T=215\mathrm{ms}$, the phase noise corresponds to a frequency stability of $1.8\times {10}^{-12}/{\tau }^{1/2}$, limited by noise in the LO distribution electronics.

Figure 4

Comparison of ZDT and conventional clock techniques. The cumulative beat-note phase is measured by comparing the noisy LO to the precision LO. The cumulative ZDT phase agrees with the cumulative beat-note phase and is bound at 440 mrad rms, whereas the cumulative single clock phase occasionally drifts by more than 10 rad.

Figure 5

Frequency uncertainty of conventional and ZDT clocks. In the first 10 seconds ZDT demonstrates $6.2\times {10}^{-12}/{\tau }^{0.91}$, characteristic of white phase noise while the single clock demonstrates $1.3\times {10}^{-11}/{\tau }^{0.57}$ due to the Dick effect.