Hyperfine structure of the $\left(3s3d\right){}^3{D}_J$ manifold of ${}^{25}\mathrm{Mg}$ i

Based on spectroscopy of the $\left(3s3p\right){}^3{P}_0\text{−}\left(3s3d\right){}^3{D}_1$ Mg i transitions for the stable isotopes ${}^{24}\mathrm{Mg}(I=0),{}^{25}\mathrm{Mg}(I=5/2)$, and ${}^{26}\mathrm{Mg}(I=0)$ we report measurements of the ${}^{25}\mathrm{Mg}\left(3s3d\right){}^3{D}_J$ hyperfine coefficients $A\left({}^3{D}_1\right)=141\pm 7,A\left({}^3{D}_2\right)=-59\pm 6$, and $A\left({}^3{D}_3\right)=-97\pm 3\mathrm{MHz}$. We find the hyperfine coefficients in agreement with state-of-the-art theoretical predictions presented here giving $A\left({}^3{D}_1\right)=143.3\pm 1.4,A\left({}^3{D}_2\right)=-48.3\pm 0.5$, and $A\left({}^3{D}_3\right)=-96\pm 1\mathrm{MHz}$. We also report measurements of the isotope shifts for the investigated transitions ${\Delta }^{24–25}=6\pm 9$ and ${\Delta }^{24–26}=59.7\pm 0.5\mathrm{MHz}$, significantly reducing the uncertainty compared to previous measurements.

Figure 1

Energy levels involved in spectroscopy of metastable magnesium $\left(3s3p\right){}^3{P}_0\text{−}\left(3s3d\right){}^3{D}_1$. Atoms are transferred to the $\left(3s3p\right){{}^3\mathrm{P}}_J$ states via electron impact. The resonance transition at 285 nm is used to determine the absolute fraction of atoms transferred to the metastable ${}^3{P}_J$ states [7, 8].

Figure 2

Schematic of the experimental system. The 383-nm light is provided by frequency doubling of a 766-nm external cavity diode laser in a bow tie cavity using a bismuth triborate (BIBO) crystal. Vertical polarization is assured by a polarizing beam splitter (PBS). A metastable magnesium beam is generated by electron impact and probed about 45 cm above the oven orifice after passing a Ø1-mm skimmer. The 383-nm light intersects the metastable beam at a right angle. Fluorescence detection is performed at right angles to the incoming laser beam and the metastable beam using a photomultiplier tube (PMT). The acousto-optic modulator (AOM) offers the possibility to perform spectroscopy with both zeroth and first orders as well as chopping the laser beam for phase-sensitive detection.

Figure 3

Fluorescence signal at 383 nm corresponding to the total ${}^3{P}_0\text{−}{}^3{D}_1$ multiplet (blue), the fitted model (dashed black, in the inset), and the 11 Voigt profiles that sum up to the fitted model. The individual Voigt profiles are the ${}^{24}\mathrm{Mg}$ transition (black), the ${}^{26}\mathrm{Mg}$ transition (red), and the nine allowed ${}^{25}\mathrm{Mg}$ transitions (green). The labels (a)–(h) correspond to the labels given in Fig. 4. The midpoint of the scanning range is used as zero for the frequency axis.

Figure 4

The fine structure levels of the ${}^{24}\mathrm{Mg}$ triplet $D$ (values from Ref. [10]) and the hfs levels of the ${}^{25}\mathrm{Mg}$ triplet $D$ (this paper). The ${}^3{D}_1$ level of ${}^{24}\mathrm{Mg}$ is used as zero for the energy axis, and the labels (a)–(h) correspond to the labels given in Fig. 3. All values are given in megahertz. Note that the number of significant figures does not correspond to the size of the error bars on the values for ${}^{25}\mathrm{Mg}$ and that the value ${\Delta }^{24–25}=6\mathrm{MHz}$ has been used to calculate the energy levels.