In-Source Laser Spectroscopy with the Laser Ion Source and Trap: First Direct Study of the Ground-State Properties of 217;219Po

D. A. Fink, T. E. Cocolios, A. N. Andreyev, S. Antalic, A. E. Barzakh, B. Bastin, D. V. Fedorov, V. N. Fedosseev, K. T. Flanagan, L. Ghys, A. Gottberg, M. Huyse, N. Imai, T. Kron, N. Lecesne, K. M. Lynch, B. A. Marsh, D. Pauwels, E. Rapisarda, S. D. Richter, R. E. Rossel, S. Rothe, M. D. Seliverstov, A. M. Sjödin, C. Van Beveren, P. Van Duppen, and K. D. A. Wendt EN Department, CERN, CH-1211 Geneva, Switzerland Max-Planck-Institut für Kernphysik, DE-69117 Heidelberg, Germany Fakultät für Physik und Astronomie, Ruprecht-Karls Universität, DE-69120 Heidelberg, Germany School of Physics and Astronomy, The University of Manchester, Manchester M13 9PL, United Kingdom ISOLDE, PH Department, CERN, CH-1211 Geneva, Switzerland Department of Physics, The University of York, York YO10 5DD, United Kingdom Advanced Science Research Centre, Japan Atomic Energy Agency, Tokai-mura, Ibaraki 319-1195, Japan Department of Nuclear Physics and Biophysics, Comenius University, SK-842 48 Bratislava, Slovakia Petersburg Nuclear Physics Institute, NRC Kurchatov Institute, 188300 Gatchina, Russia Grand Accélérateur National d’Ions Lourds, FR-14076 Caen, France KU Leuven, Instituut voor Kernen Stralingsfysica, BE-3001 Leuven, Belgium Belgian Nuclear Research Centre SCK•CEN, BE-2400 Mol, Belgium Instituto de Estructura de la Materia, CSIC, ES-28006 Madrid, Spain High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan Institut für Physik, Johannes Gutenberg-Universität Mainz, DE-55128 Mainz, Germany (Received 22 October 2014; published 20 February 2015)


I. INTRODUCTION
The study of radioactive neutron-rich nuclei in the lead region, where Z ¼ 82 is a nuclear magic number, is an experimental challenge.This region is inaccessible by fusion-evaporation reactions with beams of stable nuclei, while the cross section for fragmentation and spallation of heavier nuclei is distributed over a wide range of isotopes, leaving those of interest hidden by more abundant contaminants.A great effort has nonetheless been invested in recent years to study nuclei far beyond the nuclear magic number N ¼ 126, due to their importance for the shell evolution away from nuclear stability, for astronuclear processes, and for the search for physics beyond the standard model in octupole-deformed atomic nuclei [1][2][3][4][5].
Because of advances at on-line radioactive ion-beam facilities [6], such as ISOLDE at CERN [7], these isotopes are now routinely becoming available from proton-induced nuclear reactions with 238 U.The reaction products are extracted from the target, ionized, separated in magnetic fields, and delivered to experiments as low-energy ion beams (up to 60 keV) of a chosen isotope.This approach is referred to as isotope separation on-line (ISOL).In many cases, the isotope-separation process does not provide sufficient purity and an unwanted isobaric contaminant may dominate the extracted ion beam, thereby limiting the feasibility of a particular experiment.The introduction of the resonance laser-ionization technique has increased the element selectivity of the ionization process [8], which has enabled a large research program to be undertaken, among others, in the lead region; see, e.g., Refs.[9][10][11][12][13][14][15][16][17].Alternative production routes are found in the combination of in-flight fusion-evaporation or fragmentation facilities with a gas catcher [18][19][20].Although the underlying concepts are different, the final approach of ionization or ion extraction, followed by separation, remains similar, and those facilities will face the same difficulties as ISOL facilities.
While the resonance ionization process enhances selectively the element of interest, it does not specifically reduce the other ionization mechanisms, such as surface ionization.The Resonance Ionization Laser Ion Source (RILIS) [21] at ISOLDE makes use of a standard-surface-ionsource cavity as it provides a confined space for the laser-atom interaction.However, the elements with a low ionization energy (IE) (e.g., 87 Fr, 88 Ra) are surface ionized at the walls of the hot cavity alongside the laser-ionization process.Attempts to suppress the surface ions through the use of low-work-function cavity materials [22] or the pulsed-release technique [23] have been successfully applied in cases where only moderate selectivity enhancement is sufficient (e.g., for 215 ≤ A ≤ 218 [9][10][11][12][13]).In the case of 219 Po, however, the RILIS-ionized polonium beam is typically contaminated with 10 5 times more francium if ionization takes place inside a surface-ion-source cavity.
The Laser Ion Source and Trap (LIST) uses an alternative approach.It geometrically decouples the volume where the laser ionization takes place from where other ionization mechanisms occur.The LIST technology, proposed first in 2003 [24], was predominantly developed at the Johannes Gutenberg-Universität Mainz, consecutively undergoing a number of adaptations to match conditions and operation at on-line facilities [25,26].In a slightly modified version using a sextupole structure, it was also applied to a laser ion source coupled with a gas catcher, first at the LISOL setup of the Cyclotron Research Centre (Louvain-La-Neuve, Belgium) [27] and more recently at the IGISOL facility of the University of Jyväskylä (Finland) [28,29].Lately, it has been successfully launched as a standard add-on unit for use with the hot-cavity RILIS at the thick-target radioactive ion-beam facilities ISOLDE [30,31] and TRIUMF-ISAC [32,33].
The new developments at ISAC have enabled the successful study of radioactive magnesium isotopes with A < 24 in high-purity conditions [34,35].These studies have demonstrated a suppression factor for contaminants of 10 6 against a reduction of the beam of interest of 50.
A similar performance is found at ISOLDE in the same region, where a suppression of ≫ 10 4 against a reduction of the beam of interest of 50 is measured.In both cases, beams are extracted after the irradiation of a light target (SiC at ISAC, Ti foils at ISOLDE) with a primary proton beam.Such a combination of target and ion source cannot, however, produce the beams of interest in this discussion, for which thorium or uranium is required as the target material.Following the irradiation of such targets, much higher ion currents and radiation levels are expected, from which additional effects may arise (e.g., radiation-induced ionization, space-charge saturation).
We reported recently on the technical aspects of the first successful use of the LIST at ISOLDE for on-line production of radioisotopes from a uranium-based target with an improved laser ionization efficiency of a factor of 2.5 leading to a reduced loss of the beam of interest of only a factor of 20 [36].We present here a performance study of the LIST for nuclei in the region of Z > 82 and N > 126 and its suitability for the study of nuclear ground-state properties.In particular, the list has enabled the first measurement of the hyperfine structure and isotope shift of 217 Po and the first nuclear decay study of 219 Po, which are both of interest to delineate the region of octupoledeformed nuclei [5,37].

A. Production of polonium at ISOLDE
At ISOLDE, polonium isotopes are produced by spallation and fragmentation reactions of 238 U when the pulsed 1.4-GeV proton beam of the CERN proton synchrotron (PS) Booster impinges on the ISOLDE uranium-carbide target.The reaction products diffuse from the porous target material and leave the target container by effusion through the transfer line and hot cavity.To leave the target assembly, the atoms must effuse along a tantalum transfer line and ionizer cavity.Resistive heating is used to maintain a temperature of T ≈ 2300 K to enhance the effusion process.After ionization in the LIST (as detailed in a subsequent paragraph), the reaction products are extracted as an ion beam with an energy of up to 60 keV.Most of the results presented here are obtained using the general purpose separator (GPS) with a resolving power typically R ¼ ðm=ΔmÞ ≈ 2500.The beam is then delivered to the Windmill α-decay setup [38].A description of this process can be found in Fig. 1 and Ref. [36].

B. Resonant ionization of polonium
Reference [39] provides a technical description of the ISOLDE RILIS setup.The three-step photoexcitation scheme for polonium is shown in Fig. 2 [40].This scheme uses a frequency-tripled dye laser operating at 255.8 nm (P ≈ 20 mW) for the transition to the 6p 3 ð4S ∘ Þ7s 5 S ∘ 2 first excited state, a tunable narrow-band (NB) Ti:sapphire laser [41] at approximately 843.4 nm (P ≈ 1 W, FWHM 0.8 GHz) for the scan of the transition to the 6p 3 ð 4 S ∘ Þ7p 5 P 2 second excited state, and the 532-nm (P > 10 W) second harmonic output of a neodymiumdoped yttrium-aluminum-garnet (Nd:YAG) laser for the nonresonant transition over the ionization energy (IE) into the continuum.The output power of the NB Ti:sapphire laser for the second transition is reduced to approximately 3 mW in order to minimize spectrum broadening due to saturation of resonances.

C. LIST
The LIST is coupled downstream of the target at a distance of 2.5 mm from the exit of the ion-source hot cavity as illustrated in Fig. 1.This distance is twice shorter than in the preceding on-line test at ISOLDE and has led to an improvement of the laser-ionization efficiency by a factor of 2.5 in this work [31].
Figure 3 shows a transverse cut through the LIST.The LIST consists of two circular electrodes (repeller and extractor electrodes) with central bore diameters of 11 mm at each end of a cylindrical housing of 90-mm length and 38-mm diameter.The latter shields the enclosed radio-frequency quadrupole (RFQ) ion-guide structure from external fields such as the field of the ISOLDE extraction electrode (up to 60 kV) that is installed about 6 cm after the LIST in the ISOLDE beam line (see Fig. 1).The RFQ shield is electrically connected with the end electrode and the base of the target chamber, while the potential U rep of the front electrode (the so-called "repeller") could be controlled remotely in the range of AE500 V.
Ions inside the LIST RFQ structure are guided along the axis of the device toward the extractor of the ISOLDE mass separator by a weak longitudinal potential gradient between the repeller and LIST extractor electrode (see Fig. 3).The RFQ structure with quadrupole rods of 10-mm diameter and a free-field radius of 7.5 mm provides transverse 1.An illustration of the LIST operated at the ISOLDE mass-separator facility.The LIST is attached to the target-ion-source assembly.Protons irradiate the target, and reaction products diffuse and effuse from the target into the LIST RFQ, where the atoms are ionized by the laser radiations.After extraction and acceleration by the extractor electrode at up to 60 kV, the ions are sent through the magnetic-dipole mass separator to the Windmill detection setup and data acquisition (DAQ).FIG. 2. (a) Three-step laser-ionization scheme for polonium.(b) Close-up of the possible hyperfine levels for the nuclear spin I ¼ 9=2 with total angular momentum F and corresponding transitions (solid red lines) for the studied transition from the 7s 5 S ∘ 2 state to the 7p 5 P 2 state.The hyperfine structure is not to scale.confinement of ions as they drift through the LIST toward the extraction field.It operates with a sinusoidal rf signal of frequency f rf ¼ 1.15 MHz and an amplitude of up to V rf ¼ 500 V pp .While resonance laser ionization occurs wherever there is a laser-atom overlap (inside the hot cavity and along the axis of the LIST cavity), by adjusting the potential U rep to the repeller electrode, it is possible to select between two modes of operation.
(1) The first is the ion-guide mode (typically U rep ¼ −50 V): Ions from the hot cavity pass the repeller electrode and are guided by the RFQ to the extraction region.This mode of operation is the equivalent of the standard hot-cavity RILIS configuration.
(2) The second is the LIST mode (typically U rep ¼ þ7 to þ50 V, depending on the contaminant): Hotcavity ions are repelled, and only ions created inside the LIST structure are extracted.This mode is the high-selectivity mode, which provides a suppression of surface ions of at least 10 3 ions=s but results in a laser-ionization efficiency drop by a factor of 20, due to the relatively poor geometrical overlap of the laser and atom beams inside the RFQ.For a more detailed description of the technical aspects of the LIST and a summary of the performance during this and earlier on-line tests, we refer to Refs.[31,36].

D. The Windmill α-decay detection setup
The Windmill α-decay detection setup [38] consists of a rotatable wheel, which holds ten thin carbon foils (10-mm diameter, 20 μg=cm 2 [42]).The ion-beam implantation site is surrounded by two silicon detectors covering 51% of 4π for the detection of charged particles [43].One of the silicon detectors has an 8-mm aperture to allow the beam to pass through.The wheel can be rotated to remove longlived activity from the implantation position.A decay position is also equipped with two silicon detectors to observe the remaining decaying activity immediately after the first wheel motion.The data are collected with a triggerless acquisition system, which allows extraction of the time structure and appointment of coincidences between the detectors as well as with the laser pulses.

III. PERFORMANCE OF THE LIST A. Suppression of surface-ionized ions
The suppression of surface-ionized ions is studied with beams of radioactive francium produced on-line at ISOLDE.The francium ions cover a wide range of masses, from A ¼ 200 [44] to 233 [45].Their yields vary from < 1 to > 10 8 atoms per second, depending on the reaction cross section, the diffusion and effusion properties of the target, and the half-life of each isotope [46].In the region of A ≥ 220, the yields are 10 7 ions per second [46], while the predicted yields of the isobaric polonium atoms are < 10 3 .
In order to demonstrate the suitability of the LIST in the region of the nuclear chart with A ∼ 220, the suppression of 218 Fr has been studied.In Fig. 4, the difference in the α-decay energy spectrum between the LIST data (blue lines) and standard surface ionization (red lines) is presented.The standard-surface-ion spectrum is acquired with the same detection setup as the LIST data but with a different target unit connected through a different separator magnet (HRS) with a higher mass-resolving power of R ≈ 7000 but reduced transmission.The beam is purified in both cases using the pulsed-release technique [23], with beam-gate delays of 200 ms for the surface data and 100 ms for the LIST data, respectively.Considering the half-life of obtained in two different runs using the LIST to suppress the surface ions (blue lines) and using a standard-surface-ion source (red lines).The same measurement setup is used in both experiments.A beam-gate waiting time for the pulsed-release technique of 100 ms is applied for measurement with the LIST and 200 ms for the measurement with the surface-ion source.
Both data sets are normalized to counts per 100 s.
218 Fr (T 1=2 ¼ 22 ms and 1 ms [47]), the suppression effect from the pulsed-release technique is expected to be greater for the surface data.In spite of the smaller transmission and the pulsed-release technique, the presence of 218 Fr in the beam is significant in the surface data, as demonstrated by its characteristic α decay and that of its daughter nucleus 214 At.It can be explained by the decay of radiogenic 222 Ac (T 1=2 ¼ 63 s) inside the target, which results in the continuous production of 218 Fr, independent of the proton impact on the target.On the contrary, the LIST data show only trace amounts of 218 Fr.The suppression is applied on the ion source rather than on the timing of the production mechanism.As a consequence, the suppression power of the LIST is independent of the irradiation pattern of the target and does not differentiate between the directly produced ions and those arising from the decay of a precursor accumulated in the target material.
The effect of the repeller in both ion-guide and RILIS modes has been studied in greater detail with beams of 205;212 Fr.The ion-beam rate for each isotope as a function of the applied voltage is shown in Fig. 5.The 205 Fr data are collected with the Windmill while 212 Fr is studied with a Faraday cup.In the case of 205 Fr, the highest suppression factor from the ion-guide mode to the LIST mode is 2540.It is, however, only 70 for 212 Fr.
In Fig. 5, it is observed that the intensities of 205 Fr and 212 Fr have a different dependency on the repeller potential.The intensity of the 205 Fr ions drops rapidly for positive repeller voltages.This rapid drop is due to most of the ions coming from the hot cavity being directly repelled by the repeller potential and the few ions that are observed for positive repeller potentials being ions that are ionized closed to the repeller electrode by secondary ionization mechanisms (e.g., reionization of neutralized francium that condenses onto the repeller electrode).These ions have an offset of the kinetic energy corresponding to the potential at this position (approximately 0.6 × U rep on the longitudinal axis at the repeller position), which in turn leads to losses inside the separator magnet due to the limited energy acceptance.This dependency on the repeller potential is also observed for the laser-ionized ions, most of which are ionized close to the repeller electrode and whose intensity decreases with increased repeller potential U rep .This particular dependency is evidenced with the study of the ion time profile in Ref. [31].For 212 Fr, the intensity remains stable for repeller bias voltages from þ50 to þ500 V.
This lack of sensitivity to the bias voltages indicates that most of these ions are not ionized in the direct vicinity of the repeller electrode, whose penetration inside the LIST cavity is short but efficient.On the contrary, these ions are ionized alongside the longitudinal axis in the center of the LIST where the repeller potential and the potential of the ISOLDE extraction electrode do not penetrate, and only the initial thermal energy and the RFQ potential weak longitudinal potential gradient is guiding the ions.The significant difference in kinetic energies of laser-ionized ions and contaminants allows us to increase the selectivity in future experiments by tuning the magnetic field of the separator magnet, which has an energy acceptance of 50 eV.

B. Alternative ionization mechanisms
In order to better understand the difference in suppression of those isotopes, further studies of the ions' time profile are performed.
The time profile of the release of elements from the target material after the impact of a proton pulse is phenomenologically known to consist of a fast rise time and a long exponential decay [48].In the case of alkali elements, the release is mostly contained within the 1.2-s separation between two proton pulses irradiating the target [46,48].This release profile results in the francium beams to be structured with peaks synchronized with the proton-pulse impact on target.
The current of 212 Fr in ion-guide mode shows such a structure, but it disappears completely for a repeller voltage > þ7 V.Although the suppression is only a factor of 70, the observation of a continuous output indicates that the ions are not originating from the proton impact on target.
The alternative is that those ions are produced inside the LIST itself.Two parameters should be considered: (1) the origin of the atoms and (2) the ionization mechanism.There are two possible origins for the atoms: evaporation of condensed atoms on the rods or decay products from condensed atoms.In the case of 212 Fr, the signal observed in the Faraday cup in LIST mode is decaying with a 20-min half-life, corresponding to that of 212 Fr, while its possible precursors ( 212 Ra from β decay and 216 Ac from α decay) have much shorter half-lives (13 s [49] and 44 ms [50], respectively).The long half-life behavior shows that it is a directly evaporated atom sample and not a decay product.There are also several other possible ionization mechanisms: surface ionization off the rods, electron-impact ionization, and nonresonant laser ionization.The surface ionization is unlikely as the rods are shielded from the hot ion source and should not be sufficiently hot for surface ionization to occur.The signal is also not responding to the laser radiation.It is therefore believed to arise from electron impact following the acceleration of electrons by the rf field of the LIST quadrupoles and the field of the repeller electrode.
An example of decay-induced ion-beam production is also observed in beams of 216 At and 217 Rn, as shown in the α-decay energy spectrum of, e.g., A ¼ 216 in Fig. 6.There are no precursors observed on the foil, meaning 216 At and 217 Rn are delivered as a beam.There is also no correlation observed with the proton impact on target.Finally, their half-lives (300 μs [50] and 540 μs [51], respectively) are so short that those isotopes cannot be extracted from processes occurring inside the target, meaning that it comes directly from the LIST.
It is suggested that the parent nuclei, namely, 220 Fr and 221 Ra, condensate on the rods of the rf structure of the LIST where they α decay, leaving their daughter nucleus in an ion form.While most of those ions recoil out of the pseudopotential produced by the RFQ, some are caught and extracted with the rest of the ions and delivered to the experiment.
The production rates of 216 At and 217 Rn are 0.1 ions per second.These two isotopes represent the shortest-lived radioactive ion beams ever produced at ISOLDE since 14 Be (T 1=2 ¼ 4.35 ms) [52].
This additional source of nonresonant ions can be reduced in future designs of the LIST by producing an RFQ with thinner rods for a smaller surface area on which to condensate.Alternatively, this current design can be used to produce exotic beams otherwise not accessible at ISOL facilities.

C. Laser ionization
It has been observed in previous studies that the efficiency of the laser ionization in the LIST mode is lower than in the ion-guide mode [31,36].The losses are mostly due to the reduced spatial overlap of the atomic flux and the laser beams inside the LIST cavity in contrast to the 3-mm-diameter hot cavity that is used in normal RILIS operation or in ion-guide mode.The atomic beam divergence in the LIST cavity leads to a fast reduction of the atom density alongside the longitudinal axis as a function of the distance to the hot cavity.The ionizing lasers provide beams of 3-mm diameter, and a large fraction of the atoms is not irradiated.The interest in the LIST relies on the grounds that losses are smaller for the ions of interest than the suppression is for the contaminants.
Similarly to the case of 218 Fr, the rates of ionization of polonium in a standard laser ion-source configuration (RILIS) and with the LIST are compared in Fig. 7 smaller solid-angle coverage [40] and are corrected accordingly.Both data sets are normalized to counts per 100 s.The RILIS laser system is also different for the two measurements, which can induce differences in the ionization efficiency.In both spectra, the presence of 196 Po is clearly visible, reduced by a factor of 20 with the LIST.Alongside the study of francium suppression versus repeller potential, a reduction of laser-ionization efficiency by a factor of 5 is observed for 208 Po for the transition from ion-guide mode to LIST mode, as shown in Fig. 5.However, this number is also dependent on the proton current on target, relating to the total number of charged particles in the hot cavity, and on the size and space position of laser beams [36].In other measurements, loss factors of 10 and 20 are measured for polonium and magnesium, respectively.In total, a net improvement of a factor > 500 in beam purity (selectivity) can be reached in the cases where no alternative ionization mechanisms are present [31].

D. Resonance line profile
The linewidth and centroid of the 843.4-nm transition are systematically studied with 208 Po.The resonance spectra are recorded using Faraday-cup detection subsequently in both LIST and ion-guide modes and are shown in Fig. 8.The spectrum of the same isotope studied in the same way for the standard RILIS with a different target unit is shown for comparison.
The full width at half maximum of the single-peak resonance profiles are very similar for all three modes of operations, ranging between 2.2 and 2.5 GHz.This profile is defined by the Doppler width associated with the velocity distribution of the atoms in the ion source (source temperature and atom-beam divergence) and by the spectral density of the laser power [38].The small differences arise from drifts in the laser system over time.Those fit within the standard fluctuations of the operating parameters of the RILIS lasers and have no impact for the study of hyperfine structures and isotope shifts of radioactive isotopes.
The transition centroids are 11853.6855ð30Þcm −1 in ion-guide mode, 11853.6875ð30Þcm −1 in LIST mode, and 11853.6873ð30Þcm −1 with the standard RILIS.All measurements agree within each other's uncertainty.This agreement highlights that the different conditions of laser-atom interactions in the studied operation modes of LIST and RILIS do not introduce any systematic shift in the resonance frequency and further confirms that the LIST is a suitable setup to perform in-source laser spectroscopy.

A. Measurement
Data acquisition is performed in the same manner as the previous polonium campaigns at ISOLDE [38], measuring the α-decay rates at the Windmill detection setup while scanning the laser frequency of the 843.4-nm transition.A newly developed laser-scanning control [53] is synchronized to the PS Booster super cycle (46.8s) to ensure the reproducibility of the target irradiation at each frequency step.The pulsed-release technique [23] with a beam-gate waiting time of 100 ms after proton impact is applied to partially suppress the tail of the mass distribution of the short-lived, neighboring isobar 218 Fr (T 1=2 ¼ 22 ms).
Laser-frequency scans are recorded at mass A ¼ 217 and additionally at A ¼ 218 for the determination of the isotope shift.Because of the limited resolution of the GPS, 216 Po is observed simultaneously in the scans for mass A ¼ 217, rendering a specific reference scan for mass A ¼ 216 unnecessary.A further advantage of having both isotopes in the same mass spectra is that it allows for a direct measurement of the isotope shift between these two isotopes.In this approach, 216 Po serves as a reference isotope avoiding any laser-or acquisition-related frequency shifts [54,55].For mass A ¼ 219, a scan is performed, but insufficient statistics prevent the determination of the hyperfine structure or isotope shift.

B. α-decay energy spectra
Figure 9 shows representative α-decay energy spectra measured by the silicon detector behind the carbon foil and the mass separator set to (a) A ¼ 217 and (b) A ¼ 218 and integrated over all frequency steps of the laser scan for 216;217 Po and 218 Po, respectively.
In the case of mass A ¼ 217, the two strongest lines belong to laser-ionized 217 Po (E α ¼ 6536 keV) and 216 Po (E α ¼ 6778 keV).The next significant peak is 217 At (E α ¼ 7070 keV), which is produced by β − decay (5% Po recorded using Faraday-cup detection for both operation modes of the LIST (FWHM ∼2.5 GHz) and for a standard RILIS (FWHM ∼2.2 GHz).The lines along the data points represent the best fits to the data [38].The vertical lines highlight the centroids extracted from these fits.Apart from the decay products of the polonium isotopes and the rather weak peaks of 216 At and 217 Rn, no additional lines are visible and the polonium peaks are well isolated.A significant consequence of the LIST application is the absence of the 213;220;221 Fr α-decay lines in the α-decay energy spectra at A ¼ 217 and A ¼ 218, in contrast to the previous experiments [16,38], demonstrating the effective suppression of surface-ionized isotopes by the LIST.

C. Laser resonance spectra
Laser spectra for 216-218 Po (see Fig. 10) are obtained by integrating the counts in the respective α-decay lines seen in Fig. 9 for every laser-frequency step.The even-even isotopes 216;218 Po show the single resonance for the groundstate spin I π ¼ 0 þ and a clear isotope shift with respect to each other, while the odd-A isotope 217 Po scan exhibits a complex hyperfine structure.
The spectra are fitted using the formalism described in Ref. [38].The isotope shift δν and the hyperfine parameters A and B ( 217 Po only) are extracted.They are presented in Tables I and II.Considering the uncertainty of the groundstate spin of 217 Po [56,57], the hyperfine structure is fitted with I ¼ 9=2 and 11=2 and both results are presented in the tables.No real difference in the fit is found between the different assignments.

D. Changes in the mean-square charge radii
The changes in the mean-square charge radii δhr 2 i are extracted from the δν following the formalism presented in Refs.[16,58] using atomic parameters from large-scale atomic calculations [59].The extracted δhr 2 i are listed in Table I and are shown in Fig. 11.
The isotope shift δν 216;218 has been evaluated with both the standard RILIS configuration [16] and with the LIST. 216Po is used as the reference isotope throughout this work, providing an ideal comparison value between the two experimental setups.As given in Table I, δν 216;218 agrees within uncertainties of 1%, removing the concern that the LIST might introduce systematic effects and validating its use for isotope-shift studies.
An outstanding feature of the δhr 2 i in the region of interest is the odd-even staggering reversal already identified in 86 Rn [64], 87 Fr [37,65], and 88 Ra [66].This phenomenon has been found to be correlated to the island of octupole deformation [65,66], but the nature of the connection between the two phenomena remains under question [67].The visibility of the odd-even staggering is enhanced by removing the even-A trend from the δhr 2 i by calculating the relative odd-even staggering parameter δ: This parameter δ is shown for the isotopes of even-Z elements in Fig. 12. Odd-A isotopes with N ≤ 126 systematically show δ ≤ 0, as is characteristic across the nuclear chart.However, for N > 126, odd-even staggering reversal is found for 86 Rn and 88 Ra, with δðN ¼ 133; 135; 137Þ > 0. This phenomenon is not observed in 217 Po 133 , which staggers in the usual negative way.The absence of staggering indicates that 217 Po is located outside the region of odd-even staggering reversal.More complex parameters exist to evaluate the odd-even Changes in the mean-square charge radii δhr 2 i along the polonium isotope chain.The experimental data are shown in red [16,58,60], the prediction from the spherical droplet model (DM) in black [61,62], and mean-field calculations in blue [63].
staggering, as discussed, e.g., in Refs.[68,69], but the polonium data at hand are too limited for a conclusive interpretation.

E. Electromagnetic moments
The electromagnetic moments μ and Q S are extracted following the formalism presented in Ref. [38] using 207 Po as a reference isotope.Resulting values for the final electromagnetic moments are given in Table II, together with the g factors.
The latter data enable a comparison of different isotopes with similar nuclear configurations but different spins.The g factors associated with either the quenched free-particle g factor g S ¼ 0.6g S;free [74] or the single-particle g factor [75] for a neutron occupying the g 9=2 orbital are, respectively, −0.255 and −0.294 with I ¼ 9=2.The measured data for the two isotopes 211 Po and 217 Po, with respective g factors of −0.266ð19Þ and −0.246ð20Þ, show that in this respect, the two isotopes are very similar, both exhibiting the singleparticle value for νg 9=2 .
It is worth mentioning that the g factor for a single neutron occupying the i 11=2 orbital would be þ0.177[74] or þ0.126 [75], both of which are inconsistent with the extracted value of −0.205ð16Þ for I ¼ 11=2.This discrepancy in the case of I ¼ 11=2 further supports the I π ¼ 9=2 þ assignment, which is connected with the absence of octupole deformation [57], while the possibility of octupole deformation in the ground state of 217 Po is associated with the occupancy of the i 11=2 orbital [56].
The electric quadrupole moment for 217 Po is found to be consistent with 0. It further supports the claim that this isotope does not display any static deformation in its ground state [57].

A. Measurement
The acquisition cycles for 219 Po are determined by two parameters: the implantation period, during which the ion beam is delivered to the foil at the implantation position, and the possibly longer observation period, during which the decay is observed.Once a full cycle is completed, the activity at the implantation site is removed by turning the wheel, bringing the activated sample to the decay position and a fresh foil to the implantation position.The pulsedrelease technique [23] with a beam-gate waiting time of 100 ms after proton impact is applied to minimize isobaric  219 Fr (T 1=2 ¼ 21 ms) in addition to purification with the LIST.

B. α-decay spectra
The α-decay energy spectrum at A ¼ 219 is shown in Fig. 13 for the (a) implantation and (b) decay sites using 300-s implantation and observation periods.Known αdecay transitions from 219 At, 219 Rn, and its daughter 215 Po from α decay are identified.The reduction in the intensity ratio of the 6819-keV peak from the decay of 219 Rn with respect to 219 At between the implantation and decay positions shows that most of the 219 Rn stems directly from the ion beam, rather than from in-foil β decay of 219 At.Since 219 Rn has a short half-life (T 1=2 ¼ 3.96 s) with respect to the observation cycle (ΔT ¼ 300 s), 99.7% of the implanted 219 Rn decays within the first 12 s of the observation at the decay position, while only small amounts are steadily produced via the β decay of 219 At (T 1=2 ¼ 56 s).The intensity of the peak at 7386 keV, stemming from the α decay of 215 Po, depends on both the α decay of 219 Rn and the β decay of 215 Bi, with a longer half-life (T 1=2 ¼ 7.7 min).Decays from neighboring masses are also seen: 218 Rn, its daughter 214 Po, 220;221 Fr, and their daughters 216;217 At.The presence of directly ionized radon and francium originates from the processes described in the section on alternative ionization mechanisms.Finally, long-lived 222 Rn (T 1=2 ¼ 3.825 days) is also seen in the spectrum at 5489 keV, arising from one of the carbon foil, which has been contaminated in the course of the experimental campaign.
By comparing the spectra with and without laser irradiation, it is possible to identify the α-decay lines related to polonium; see Fig. 14.The peaks observed in the lasers-off-subtracted spectrum are identified as arising from the α decay of 219 Po and its progeny 219 At, 219 Rn, and 215 Po.As 219 Rn is also produced directly, the subtraction of two large peaks occurring in both spectra results in large fluctuations of the counts around 6819 and 7386 keV ( 215 Po).
The properties of those four isotopes are summarized in Table III.The α-decay energies for 219 Po and 219 At are determined to be 5806(5) and 6228(5) keV, respectively.

C. Branching ratios
Using collections made with 300 s for both implantation and observation periods, it is possible to deduce the branching ratios b α and b β of α and β decays, respectively, in the case of 219 Po and 219 At.As seen in Fig. 15, the branching ratios of these isotopes can be determined by comparing the number of direct α decays with the number of α decays from their daughter nuclei via β decay.
For 219 At, the numbers of events at 6228 keV (α decay) and 6819 keV (β decay) are compared with each other, including the known b α ð 219 Rn; 6819 keVÞ ¼ 79.4ð10Þ% [76].Only data from the decay position are considered, so that the directly implanted 219 Rn ions can be minimized and their influence disregarded.Recoils out of the carbon foil following β decay are negligible.A branching ratio of b α ð 219 AtÞ ¼ 93.6ð10Þ% is found.It is in reasonable agreement with the previous estimate of approximately 97% [78] that was determined in the study of the α decay of natural 227 Ac samples.
The same procedure is applied to 219 Po by comparing the number of events at the implantation position at 5806 keV  Half-life taken from Ref. [76].
(α decay) and 6228 keV (β decay), taking into account the new branching ratios that have been determined for the decay of 219 At.Since no 219 At is directly implanted on the foils, the full data may be utilized.Once again, the β-decay recoils are ignored.A branching ratio of b α ð 219 PoÞ ¼ 28.2ð20Þ% is found.The results are summarized in Table III.

D. Half-life
In the course of the experiment, a 1200-s decay-only period is recorded at the decay site.The half-life of 219 Po is determined to be 620 (59) s by fitting the time behavior of the α-decay peak at 5806 keV, shown in Fig. 16, to a single exponential decay curve.

E. Hindrance factors
The α-decay hindrance factors (HFs) are calculated following the formalism of Rasmussen [79] by comparing to the α decay of 218 Po [partial decay width 117(1) keV].Values of HFð 219 PoÞ ¼ 1.5 and HFð 219 AtÞ ¼ 1.1 are determined.Both isotopes exhibit, therefore, unhindered α decay, which indicates that their nuclear ground-state spin and configuration are similar to their daughter nuclei 215 Pb and 215 Bi, respectively.In the case of 219 Po, this configuration would correspond to a spin I π ¼ ð9=2 þ Þ arising from a valence neutron in the ν2g 9=2 orbital [12].This configuration is consistent with the predictions of the spherical shell model of the nucleus and suggests that 219 Po does not display any sign of octupole deformation.In the case of 219 At, the spin is I π ¼ ð9=2 − Þ, arising from a valence proton in the π1h 9=2 orbital.The complete information on the decay of 219 Po and 219 At is shown in Fig. 17.

VI. CONCLUSIONS
A substantial improvement of the purity of polonium ion beams produced with RILIS has been obtained due to implementation of the LIST.Experiments on in-source laser spectroscopy of 216-218 Po and nuclear decay study of 219 Po have been performed using the LIST.This study represents the first use of the device for a dedicated experiment at CERN ISOLDE and its first use for on-line hyperfine structure and decay studies of radioisotopes.The comparison of the laser spectroscopy data with previously measured 216;218 Po shows that no systematic effects were added by the use of the LIST.Furthermore, the enhanced selectivity of the device has enabled the study of isotopes with A > 218 for the first time.This achievement opens up a new series of experiments in this region of the nuclear chart, with further experiments already planned on neutronrich 81 Tl and 84 Po.
The systematic study of the δhr 2 i odd-even staggering and of the nuclear electromagnetic moments indicates that 217 Po is located outside the region of reflection asymmetry, which would be associated with the region of odd-even staggering reversal.The study of more neutron-rich isotopes of polonium and of the neutron-rich 85 At isotopes is, however, required to fully delineate those regions of interest.The combination of the LIST for laser spectroscopy with the ISOLTRAP multireflection time-of-flight mass spectrometer for efficient identification of beam  III) and Ref. [76].
composition and ion counting [39,80] will render these investigations possible.Finally, the nuclear decay properties of 219 Po have been determined for the first time while knowledge on those for 219 At has been improved.The properties of 219 Po are consistent with the systematics of that region of the nuclear chart, for which the structure of the odd-mass polonium isotopes is dominated by a spherical neutron g 9=2 orbital.

FIG. 3 .FIG. 4 .
FIG.3.A transverse cut through the LIST with its most important parts.Insulators are colored in light gray.

FIG. 5 .
FIG.5.Effect of the repeller potential on the ion current of laser-ionized polonium and surface-ionized francium isotopes.Negative voltages correspond to the ion-guide mode, while positive voltages correspond to the LIST mode.Data taken with different techniques are normalized to ions per second, and the detection limit for the Faraday-cup measurements is shown with the dotted line.

FIG. 8 .
FIG.8.Resonance line profile of the transition at 843.4 nm in208 Po recorded using Faraday-cup detection for both operation modes of the LIST (FWHM ∼2.5 GHz) and for a standard RILIS (FWHM ∼2.2 GHz).The lines along the data points represent the best fits to the data[38].The vertical lines highlight the centroids extracted from these fits.

FIG. 9 .FIG. 10 .
FIG. 9. α-decay energy spectra for (a) A ¼ 217 and (b) A ¼ 218 obtained by the silicon detector behind the implantation point and integrated over all frequency steps of the laser scan for 216;217 Po and 218 Po, respectively.

FIG. 13
FIG. 13. α-decay energy spectrum for A ¼ 219 at the Windmill (a) implantation and (b) decay positions using 300-s implantation and observation periods with the lasers on polonium.The α-decay energies are in keV.Decays originating from the implantation of 219 Po are shown in red.

FIG. 15 .FIG. 16 .
FIG. 15.Decay chain starting from Po down to 211 Pb. αdecaying isotopes are shown in yellow, β-decaying isotopes in blue, and possible decay paths are highlighted with an arrow.
.196Po is laser ionized, and its α-decay energy spectrum is recorded with the Windmill.The RILIS data are collected with a different configuration of the Windmill setup with [40]ak arising from the decay of 216 At (T 1=2 ¼ 300 μs[50]) is highlighted in red.The peaks labeled in blue are part of the decay chain of 216 Po, which is partially enhanced with 255-nm laser light; those labeled in black are coming from contamination on the foils from earlier collections.lines)andwith a standard RILIS ion source (red lines). Theandard RILIS ions are studied with an earlier version of the experimental setup[40].Both data sets are normalized to counts per 100 s and to the solid-angle coverage of the current version of the Windmill setup.

TABLE I .
[16]ope shifts δν A;216 and changes in the mean-square charge radii δhr 2 i of 216;217;218 Po.The results for 216;218 Po from previous RILIS measurements[16]are shown for comparison and to calculate δν A;210 .Statistical uncertainties are given in parentheses, and systematic uncertainties originating from the atomic parameters are given in the curly brackets.For 217 Po, the results for fitting with both I ¼ 9=2 and 11=2 are presented, showing no significant difference.

TABLE II .
[38]rfine parameters, electromagnetic moments, and g factors of 207;211;217 Po from Ref.[38](and references therein) and this work, respectively.Statistical uncertainties are given in parentheses, and the total uncertainties including systematic uncertainties originating from the reference isotope 207 Po are given in the curly brackets.For 217 Po, the results for fitting with both I ¼ 9=2 and 11=2 are presented.
Difference between the α-decay energy spectrum for mass A ¼ 219 at the implantation position with lasers tuned to a resonance of polonium and lasers off, using 100-s implantation and 180-s observation times.The peak marked with an asterisk is 222 Rn, a long-lived contaminant implanted on the laser-on carbon foil (see the text for details).The subtraction of two large peaks for 219 Rn and 215 Po, arising from the direct production of 219 Rn, results in high fluctuations in the counts around 6819 and 7386 keV.

TABLE III .
Decay properties of 219 Po and its progenies 219 At, 219 Rn, and 215 Po. a