$2S$ hyperfine structure of atomic deuterium

We have measured the frequency splitting between the $(2S,F=1∕2)$ and $(2S,F=3∕2)$ hyperfine sublevels in atomic deuterium by an optical differential method based on two-photon Doppler-free spectroscopy on a cold atomic beam. The result $f_{\mathrm{HFS}}^{\left(D\right)}\left(2S\right)=40924454\left(7\right)\mathrm{Hz}$ is the most precise value for this interval to date. In comparison to previous radio-frequency measurements we have improved the accuracy by a factor of 3. The specific combination $D_{21}=8f_{\mathrm{HFS}}^{\left(D\right)}\left(2S\right)-f_{\mathrm{HFS}}^{\left(D\right)}\left(1S\right)$ of metastable and ground state hyperfine frequency intervals in deuterium derived from our measurement agrees well with the value for $D_{21}$ calculated from quantum electrodynamics.

Figure 1

Laser system for $1S\text{−}2S$ deuterium spectroscopy.

Figure 2

Vacuum part of the experiment. HV and UHV are the high- and ultrahigh vacuum regions, while $V$ is the bypass valve reducing the atomic flow escaping from the nozzle.

Figure 3

Simultaneous recording of the two transition lines (top). AOM frequencies $f_{\mathrm{AOM}}$ corresponding to the centers of the doublet and quadruplet transition lines fitted by Lorentzian functions. The frequency change is due to the drift of the reference ULE cavity (bottom).

Figure 4

Two-photon transitions between hyperfine components of the $1S$ and $2S$ levels in atomic deuterium. The spliting of magnetic sublevels in a magnetic field is also presented (not to scale).

Figure 5

The $2S$ HFS frequency vs deuterium flow. One unit corresponds to $3.6\times {10}^{17}$ particles coming to the nozzle per second. The data are obtained using the simultaneous recording method and correspond to the delay time of $\tau =410\mathrm{\mu s}$. To extrapolate to zero flow (gray point) we fitted the data with a linear function (solid line).

Figure 6

Quadruplet transition lines and corresponding Lorentzian fits for the different delays $\tau$. Although the lines detected for $\tau =10$ and 210 μs are visibly asymmetrical, the line asymmetry decreases at higher delays.

Figure 7

Experimental and theoretical values for $D_{21}$ in deuterium. The dashed lines represent the uncertainty of the theoretical value.