High-Precision Ramsey-Comb Spectroscopy at Deep Ultraviolet Wavelengths

High-precision spectroscopy in systems such as molecular hydrogen and helium ions is very interesting in view of tests of quantum electrodynamics and the proton radius puzzle. However, the required deep ultraviolet and shorter wavelengths pose serious experimental challenges. Here we show Ramsey-comb spectroscopy in the deep ultraviolet for the first time, thereby demonstrating its enabling capabilities for precision spectroscopy at short wavelengths. We excite ${}^{84}\mathrm{Kr}$ in an atomic beam on the two-photon $4p^6\to 4p^{5}5p[1/2{]}_0$ transition at 212.55 nm. It is shown that the ac-Stark shift is effectively eliminated, and combined with a counterpropagating excitation geometry to suppress Doppler effects, a transition frequency of 2 820 833 101 679(103) kHz is found. The uncertainty of our measurement is 34 times smaller than the best previous measurement, and only limited by the 27 ns lifetime of the excited state.

Figure 1

Schematic overview of the setup. Frequency-comb (FC) pulse pairs are amplified in a noncollinear optical parametric chirped-pulse amplifier (NOPCPA). Spectral interferometry is used to measure the differential phase between the amplified and original frequency-comb pulses. The amplified pulse pairs are up-converted sequentially using three BBO crystals, and the two insets show the spectral and temporal evolution of each Ramsey comb pulse, starting with the original near-infrared spectrum and ending with a “red-blue” split pulse in the deep UV near ${\lambda }_{\mathrm{tr}}=212.55\mathrm{nm}$. The UV beams are aligned exactly counterpropagating, using the fringes observed at the output of the metallic 50% beam splitter (BS), which acts as a Sagnac interferometer (SI). After excitation, the krypton atoms are ionized and extracted upwards in a time-of-flight (TOF) drift tube and detected with a channel-electron multiplier (CEM, EDR model from Dr. Sjuts Optotechnik GmbH).

Figure 2

Example of measured Ramsey signals at $\mathrm{\Delta }N=2$ and $\mathrm{\Delta }N=7$. At each $\mathrm{\Delta }N$ the delay time is scanned over $\sim 700$ attoseconds (as) and is offset by $\mathrm{\Delta }N{T}_{\text{rep}}$ [see Eq. (2)] corresponding to 15.79 and 55.28 ns, respectively, as indicated in the figure. Each data point is an average over 350 laser shots and the variation within this set is used to determine the statistical uncertainty of the data points. The signal is normalized to the maximum signal of the first Ramsey fringe. The red line is a fit to the data based on Eq. (1).

Figure 3

Measurement results of the $4p^6\to 4p^{5}5p[1/2{]}_0$ transition frequency in ${}^{84}\mathrm{Kr}$. Each data point is an average over 10–20 Ramsey measurements at macrodelays $\mathrm{\Delta }N=2$ and $\mathrm{\Delta }N=7$, and the green band shows the $1\sigma$ uncertainty of the average. The first-order Doppler shift is determined by measuring the transition frequency difference for pure Kr with respect to Kr:Ne and Kr:He mixtures. The color indicates which noble gas was used to speed up the supersonic expansion, and the numeral indicates which collision point was used.

Figure 4

(a) Results of the differential phase measurement on the fundamental pulses at 850 nm between the $\mathrm{\Delta }N=2$ and $\mathrm{\Delta }N=7$ pulse pairs. The blue and red part of the amplified pulse spectrum are measured separately from which the mean shift is calculated and shown in the graph. No significant phase shift is observed within an uncertainty of 2.1 mrad (indicated with the green band), corresponding to 35 kHz on the transition frequency. (b) Results of the ac-Stark shift determination. Each data point represents an average of 10–20 measurements. Within the uncertainty of 72 kHz (indicated by the green band), no shift due to the excitation light field is observed.