In-source laser spectroscopy of dysprosium isotopes at the ISOLDE-RILIS

A number of radiogenically produced dysprosium isotopes have been studied by in-source laser spectroscopy at ISOLDE using the Resonance Ionization Laser Ion Source (RILIS). Isotope shifts were measured relative to 152 Dy in the f s 10 8 (gs) 1) o 8 ( 418.8 nm vac ) resonance transition. The electronic factor, F , and mass shift factor, M , were extracted and used for determining the changes in mean-squared charge radii for 145m Dy and 147m Dy for the irst time.


Introduction
The Resonance Ionization Laser Ion Source (RILIS) is the most selective of all ion sources available at the ISOLDE radioactive beam facility [1]. The selectivity is an intrinsic property of the ionization mechanism, based on stepwise resonance excitation and ionization via element-speciic atomic levels. The isotope production takes place inside a thick target, on which protons, provided by CERN's Proton Synchrotron Booster (PSB), impinge with an energy of 1.4 GeV. The reaction products are released from the target material and efuse via a transfer line into a resistively heated tubular cavity, where the atomlaser interaction takes place. The resulting ions are then extracted, accelerated up to 60 keV and mass separated by a dipole magnet according to their mass-to-charge ratio.
During so-called 'in-source laser spectroscopy', the RILIS lasers are used a to probe a speciic spectroscopic transition of the ionization scheme of diferent isotopes of one element. By determining the isotope shift (IS) of a chosen transition, changes in the nuclear mean-squared charge radii can be deduced. For states with nonzero nuclear spin I which exhibit a suiciently large hyperine structure (HFS), the nuclear moments (spin, magnetic dipole and electric quadrupole moments) can be extracted. Additionally, if the HFS of diferent isomers can be resolved (due to diferent spins and magnetic moments), isomer-selective ionization is possible. The spectral resolution of in-source measurements is limited by Doppler-broadening of the spectral lines inside the ion source (which is typically heated to 2100°C). There have been several experimental campaigns, in which this in-source spectroscopy has been successfully applied (e.g. [2]) or where isomer separation was provided for higher selectivity during nuclear spectroscopy experiments (e.g. [3]).
Here, we report on the irst in-source spectroscopy study of dysprosium radioisotopes, demonstrating the suitability of this method for a future extended study of IS in the dysprosium isotopic chain.

Experimental setup
The experiment was performed using beam provided by target #655 (target with tantalum rolls from mixed 25 and 6 μm foils at 1950°C with a tungsten surface ion source at 1985°C). No stable supply of dysprosium was available initially, so that the optimization was performed on radiogenically produced 159 Dy. During the experiment, a proton current of 0.2 µA was used on target, providing a continuous supply of dysprosium.
The transition chosen for the spectroscopy leads from the f s [Xe]4 6 10 10 excited level, despite the photon energy at 532 nm being below that required to reach the ionization continuum. The ionization eiciency saturates with an estimated 7 W of laser power in the ionization region ( 3 mm laser beam diameter). From this we conclude that the 532 nm light is coincidentally resonant with a second step transition to a high-lying level, from which a second 532 nm photon induces ionization via an auto-ionizing state. In the transition metals, the atomic level density, and the richness of the autoionizing spectrum, greatly increase the likelihood of such a coincidence in required transition wavelengths.
A newly developed narrow-linewidth intra-cavity frequency-doubled mode for the Ti:sapphire grating laser was applied for the irst time, scanning across the 418.8 nm transition. It will be described in more detail in [5]. The wavelength was recorded with two High-Finesse/Angstrom WS7 wavelength meters installed in the RILIS laboratory. The wavemeters were calibrated before the measurements with a CW diode laser locked to the rubidium hyperine structure. As the transition probability lies at = × A 1.26 10 8 s −1 [6], the power of the irst-step laser beam had to be reduced signiicantly, to < 1 mW in order to avoid saturation.
For the cases of Dy the ISOLDE tape station gamma detector was utilized (for more details see [7]). An overview over the yields measured with the tape station -counting is given in Table 1.

Results
In the case of Dy -decay at 639 keV was observed. The missing observation of other lines is attributed to the high background from the -decay of the surface ionized isobars. Correspondingly, only results for the isomeric state were obtained.
The optical spectra are summarized in Fig. 1  it is not possible to give separate yields, as the ratio of the isomer production is not known. For Dy 149 the isomer contribution is assumed to be negligible due to the much shorter T 1/2 . The accuracy of the measured yields can be estimated as a factor of two, taking into account the daughter activity contribution, isomer mixture and the possible contributions from the adjacent masses.  isotopes. For the odd-A dysprosium isotopes, the resolution was not high enough to resolve the HFS and therefore only the isotope shifts were extracted. The shift in center of gravity (CoG), introduced by the underlying HFS for these cases, was taken into account by estimating it with reasonable A and B HFS-constants: A-and B-constants ratios for the excited and ground states were taken from [9], for Dy m 159,148,147 the known Q and µ values [10] were used to calculate A and B constants by the standard scaling relation [11], µ ( Dy with the same shell-model coniguration ( h 11/2 ). The shift proved to be less than 40 MHz and was added to the uncertainties of the IS for the odd Dy isotopes.
were used for comparison. A 'standard' King-plot procedure (see e.g. [11]) is not possible, due to missing IS data for the light dysprosium isotopes (only r 2 are cited in [13,11]). A modiied approach was used. Starting with the well-known relation that it follows that the modiied IS is linearly dependent on the modiied r 2 with the slope equal to the electronic factor F and the intercept equal to the mass-shift constant M: The nuclear masses A and A 0 used in the calculations by Eqs. (1)-(3) were taken from [14].
As shown in Fig. 2, all newly measured modiied IS for the 418.8 nm transition, as well as the previously measured 418.8 nm for were derived for the irst time (see Table 2). . The shell efect in the r 2 (kink at = N 82) is evident for odd-and even-N isotopes. It was found previously that there is a marked isomer shift between + 1/2 ground states and 11/2 isomers in Sm 62 and Gd 64 nuclei at < N 82 [16,17]. This isomer shift leads to the disappearance of the odd-even staggering (OES) in r 2 of the 11/2 isomers. The results obtained in the present work for 11/2 dysprosium isomers do not contradict this observation, although no deinite conclusion can be inferred due to the large experimental uncertainties.

Outlook
In order to better investigate the disappearance of normal OES in the vicinity of = N 82 for the high-spin isomers in dysprosium, expected to be inluenced by the h 11/2 -state, further studies with dedicated beam time are necessary. A better resolution of the -spectra and detection eiciency, using the ISOLDE Decay Station (IDS), would help to separate the ground from isomeric state. The relative uncertainties could be additionally reduced by using the transition to the = f s p J 4 6 6 , 9 10 state at 23736.61 cm −1 ( 421.3 nm). This transition has been shown to have an isotope-shift sensitivity twice the size of the 418.8 nm transition [11]. As seen in Table 2 and Fig. 1, at the signal to background ratio larger than 1 and suicient statistics the uncertainty of the IS determination can be reduced to 100 MHz ( Dy 158 ) and lower (taking into account reduction of the uncertainty also for CoG measurement for the reference isotope). This accuracy is expected to be suicient to investigate the evolution of OES (see results for similar 11/2-state in Sm [16]). However, more accurate results may be achieved with better resolution which would enable reliable analysis of the odd Dy isotope HFS.
It is estimated that dysprosium isotopes down to around = A 141 are accessible for IS measurements by the in-source spectroscopy method, provided suicient suppression of isobaric background is achieved (e.g. with the Laser Ion Source and Trap (LIST) [18]). It is worth to note that dysprosium isotopes with < A 146 have noticeable delayed proton branching and photo-ion current monitoring by delayed protons detection might give more favorable background conditions. Near this point, a strong onset of deformation is expected which would be relected in the IS values.

Table 2
Isotope shifts and changes in the mean-square charge radii for Dy isotopes. The errors result from the itting procedure, described in Fig. 1 (9) a Present work. b Reference [13]. and data taken from [13].