Demonstration of a Timescale Based on a Stable Optical Carrier

We report on the first timescale based entirely on optical technology. Existing timescales, including those incorporating optical frequency standards, rely exclusively on microwave local oscillators owing to the lack of an optical oscillator with the required frequency predictability and stability for reliable steering. We combine a cryogenic silicon cavity exhibiting improved long-term stability and an accurate ${}^{87}\mathrm{Sr}$ lattice clock to form a timescale that outperforms them all. Our timescale accumulates an estimated time error of only $48\pm 94\mathrm{ps}$ over 34 days of operation. Our analysis indicates that this timescale is capable of reaching a stability below $1\times {10}^{-17}$ after a few months of averaging, making timekeeping at the ${10}^{-18}$ level a realistic prospect.

Figure 1

(a) An array of three lasers are locked to ultrastable Fabry-Perot resonators. A femtosecond frequency comb transfers the stability of the OLO (124 K Si cavity) from 1542 nm to a prestabilized laser at 698 nm used to perform clock spectroscopy in a one-dimensional (1D) ${}^{87}\mathrm{Sr}$ lattice clock. (b) AT1, a free running microwave timescale at NIST, is compared continuously against the OLO signal over a fiber-optic link using a hydrogen maser (ST14) as a transfer oscillator. An optical fiber link between JILA and NIST allows for stable transfer of the optical timescale to NIST for future integration into UTC (NIST) and for broadcast to the International Bureau of Weights and Measures (BIPM).

Figure 2

(a) The OLO frequency (Si) is measured at 698 nm using a ${}^{87}\mathrm{Sr}$ lattice clock. A linear plus exponential trend, $a+bt+c{e}^{-(t/d)}$, agrees well with the raw frequency data. The fit parameters are $a=24.16\mathrm{Hz}$, $b=-9.632\mathrm{Hz}/\mathrm{day}$, $c=-23.17\mathrm{Hz}$, and $d=7.813\mathrm{days}$. (b) The residuals of the OLO comparisons against the ${}^{87}\mathrm{Sr}$ clock and the NIST AT1 timescale after subtracting the drift trend from (a) from both datasets.

Figure 3

An estimate of the time error evolution of the optical timescale over the 34 day data campaign results in an integrated value of $48\pm 94\mathrm{ps}$. The peak-to-peak value of 197 ps is dominated by a 4 day window that includes the two days when the ${}^{87}\mathrm{Sr}$ clock was not operated. The rms spread in time error for two timescales based on repeated simulations of a maser steered to either a microwave or optical frequency standard are shown for comparison [21].

Figure 4

The silicon cavity stability is computed from the detrended ${}^{87}\mathrm{Sr}$ data in Fig. 2 using a gap-tolerant Allan variance similar to Ref. [36]. The data are fit to a noise model with an instability of $\sigma =1.3\times {10}^{-18}\sqrt{\tau (\mathrm{s})}$ at long averaging times. The long-term stability of the OLO is also inferred from a continuous measurement against the NIST AT1 timescale.

Figure 5

Expected fractional frequency stability of the optical timescale. The stability of our optical timescale is analyzed for two optical clock duty cycles. Our optical timescale is compared to a hydrogen maser based timescale steered to an optical lattice clock with identical uptime or a cesium fountain clock operating continuously.