*Article* **Advancing Radiation-Detected Resonance Ionization towards Heavier Elements and More Exotic Nuclides**

**Jessica Warbinek 1,2,\*, Brankica Andeli´ ¯ c 1,3,4, Michael Block 1,2,4, Premaditya Chhetri 1,4, Arno Claessens 5, Rafael Ferrer 5, Francesca Giacoppo 1,4, Oliver Kaleja 1,6, Tom Kieck 1,4, EunKang Kim 2, Mustapha Laatiaoui 2, Jeremy Lantis 2, Andrew Mistry 1,7, Danny Münzberg 1,2,4, Steven Nothhelfer 1,2,4, Sebastian Raeder 1,4, Emmanuel Rey-Herme 8, Elisabeth Rickert 1,2,4, Jekabs Romans 5, Elisa Romero-Romero 2, Marine Vandebrouck 8, Piet Van Duppen <sup>5</sup> and Thomas Walther <sup>9</sup>**


**Abstract:** RAdiation-Detected Resonance Ionization Spectroscopy (RADRIS) is a versatile method for highly sensitive laser spectroscopy studies of the heaviest actinides. Most of these nuclides need to be produced at accelerator facilities in fusion-evaporation reactions and are studied immediately after their production and separation from the primary beam due to their short half-lives and low production rates of only a few atoms per second or less. Only recently, the first laser spectroscopic investigation of nobelium (*Z* = 102) was performed by applying the RADRIS technique in a buffergas-filled stopping cell at the GSI in Darmstadt, Germany. To expand this technique to other nobelium isotopes and for the search for atomic levels in the heaviest actinide element, lawrencium (*Z* = 103), the sensitivity of the RADRIS setup needed to be further improved. Therefore, a new movable double-detector setup was developed, which enhances the overall efficiency by approximately 65 % compared to the previously used single-detector setup. Further development work was performed to enable the study of longer-lived (t1/2 > 1 h) and shorter-lived nuclides (t1/2 < 1 s) with the RADRIS method. With a new rotatable multi-detector design, the long-lived isotope 254Fm (t1/2 = 3.2 h) becomes within reach for laser spectroscopy. Upcoming experiments will also tackle the short-lived isotope 251No (t1/2 = 0.8 s) by applying a newly implemented short RADRIS measurement cycle.

**Keywords:** laser spectroscopy; resonance ionization; atomic level scheme; gas cell; radiation detection; heavy actinides

**Citation:** Warbinek, J.; Andeli´ ¯ c, B.; Block, M.; Chhetri, P.; Claessens, A.; Ferrer, R.; Giacoppo, F.; Kaleja, O.; Kieck, T.; Kim, E.; et al. Advancing Radiation-Detected Resonance Ionization towards Heavier Elements and More Exotic Nuclides. *Atoms* **2022**, *10*, 41. https://doi.org/ 10.3390/atoms10020041

Academic Editor: Alexander Kramida

Received: 28 March 2022 Accepted: 13 April 2022 Published: 21 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

The study of the heaviest elements of the actinide series has recently gained much interest in the field of modern laser-based physics research [1]. Relativistic effects strongly influence the electronic configurations of these exotic elements, altering their atomic and chemical properties. Laser spectroscopy constitutes a powerful tool to study these effects, for example by measuring the ionization potential (IP) or by probing the atomic-level structure and optical transitions. Experimental data can benchmark theoretical predictions obtained through many-body methods such as relativistic coupled-cluster (RCC), the multiconfiguration Dirac–Fock (MCDF), and the configuration interaction (CI) calculations [2]. In addition, laser spectroscopy can give access to nuclear structure observables, e.g., nuclear spin and nuclear moments or changes in the nuclear mean square charge radii, to validate and guide theoretical studies as for instance in the region of the heaviest elements around the *N* = 152 neutron shell closure [3].

To study these heavy elements, synthetic production of respective atoms is necessary as they do not naturally occur on earth. While the actinides up to fermium (*Z* = 100) can still be produced in macroscopic sample sizes from breeding processes in nuclear reactors [4], the transfermium elements (*Z* > 100) are exclusively produced in atom-at-a-time quantities in fusion-evaporation reactions at large-scale accelerator facilities. Therefore, measurements of their atomic properties are very challenging. Some of these properties can, for instance, be determined in gas- and liquid-phase chemistry experiments [5,6]. Only recently, the IPs of the heaviest actinides, ranging from fermium to lawrencium (*Z* = 103), were determined via a surface-ionization technique to meV precision [7,8]. However, the present accuracy of the IPs is limited by the applied technique, whereas the determination by laser spectroscopy can result in orders-of-magnitude higher precision. As of today, the heaviest element investigated by means of laser spectroscopy is the actinide element nobelium (*Z* = 102) [9]. The first ionization potential of this element was determined with a 50 μeV accuracy [10] to benchmark atomic theory and to probe relativistic effects on this property in the range of the heaviest elements.

The study of transfermium elements via common laser spectroscopy techniques such as collinear laser spectroscopy [11] or the hot-cavity technique [12] is often unfeasible. Due to the low production yield of only a few nuclei per second at most and the short half-lives, the application of laser spectroscopy requires a fast and extremely sensitive probing of the produced particles directly after their separation from the primary beam. The RAdiation-Detected Resonance Ionization Spectroscopy (RADRIS) method [13,14] is dedicated to the study of exotic, heavy elements, and was successfully applied to find a ground-state transition in nobelium [9].

To date, an isotopic chain of nobelium isotopes ranging from mass numbers 252 to 254 has been investigated by this method [15]. For the study of short-lived nuclides with half-lives of t1/2 < 1 s such as the isotope 251No (t1/2 = 0.8 s), additional improvements are required due to significant decay losses expected in the measurement scheme. Further limitations appear for nuclides with half-lives of t1/2 > 1 h. Moreover, a gain in the RADRIS efficiency could in general be decisive to identify resonance signals in future experiments such as for the search of yet experimentally undetermined atomic levels in elements heavier than nobelium.

Recent developments address these current limitations in terms of the range of accessible nuclides and the overall RADRIS efficiency. Here, the latest advances towards laser spectroscopy of longer-lived and shorter-lived transfermium nuclides are discussed, and corresponding results from on-line test experiments are presented. Additional development work towards an enhanced efficiency of RADRIS for the search for atomic levels in the heaviest actinide, lawrencium, will also be outlined.

#### **2. Experimental Setup**

#### *2.1. RADRIS Technique*

The RADRIS experimental setup consists of a stopping cell attached to the velocity filter SHIP at the GSI in Darmstadt [13,16]. A schematic drawing of the RADRIS setup is shown in Figure 1a. Recoil nuclei transmitted through the velocity filter enter the gas cell through an entrance window of 3.5 μm thin aluminum-coated mylar foil, supported by a stainless steel grid, and are stopped in 90 mbar argon buffer gas. A 1 mm × 25 μm Hf-strip filament positioned opposite the entrance window and biased with an attractive voltage allows collecting incoming ions which adsorb and neutralize on the filament surface. By resistive pulse-heating of the filament, collected fusion products are re-evaporated to form a cloud of neutral atoms in the vicinity of the filament. Here, the evaporation temperature is critical for an efficient desorption paired with a minimum ion background from surface ionization processes. For an optimal filament choice, the IP of the collected atom species, the work function of the filament material, and the filament temperature are key parameters to be considered. A more detailed description of the desorption from filaments can be found in [17].

After the successful evaporation from the filament, the created neutral atoms are illuminated with two lasers following a two-step resonance ionization spectroscopy (RIS) scheme. Here, a UV-pumped dye laser supplies the first excitation step while the second, non-resonant step for ionization is provided by a high-power excimer or Nd:YAG laser. Resulting laser ions are then guided via suitable electric fields towards the detection area where they are collected on a 200 nm thin aluminized kapton foil in front of a Passivated Implanted Planar Silicon (PIPS) detector. Finally, the laser-ionized fusion products are detected via their alpha-decay energy. In this way, registered signals can additionally be gated by their characteristic decay energy to discriminate contributions of decay signals from background ions and subsequent decays of daughter nuclides.

**Figure 1.** (**a**) Schematic drawing of the RADRIS setup. Incoming fusion products (green) enter the gas cell through the entrance window and thermalize in the 90 mbar argon buffer-gas environment. The recoils are collected on a Hf-strip filament, where they adsorb and neutralize. After desorption by pulse-heating the filament, an atomic cloud (blue) is formed around the filament which is illuminated with two lasers following a two-step RIS scheme. Resulting laser ions (blue) are guided towards a silicon detector by applying suitable transport electrode potentials where they are detected via their characteristic alpha-decay energy. (**b**) Previously applied RADRIS cycle for resonance ionization spectroscopy of 155Yb. Delays between blocking the beam and changing the potentials, as well as applying the filament heat-pulse, are chosen to allow settling down of the stopped recoil ions and complete switching of the voltages, respectively [13].

#### *2.2. RADRIS Measurement Cycle*

The application of the RADRIS technique is cyclic, where each cycle is divided into the filament accumulation mode and the ionization/guiding mode. In the accumulation mode, fusion products enter the gas cell and are collected on the filament by applying an attractive potential compared to the surrounding electrodes and the chamber itself [13]. After the accumulation is completed, the setup is switched to the ionization/guiding mode in which the incoming primary beam is stopped, the laser shutters are opened to expose the stopping volume to laser light, and the potentials are set such that the created laser ions are guided towards the silicon detector. To change between the two modes, electrostatic potentials applied to the chamber, the filament, and the transport electrodes surrounding the filament need to be switched by giving an analog trigger to the power supplies after the primary beam is stopped. A time of approximately 0.3 s is required for the potentials to reach the set values. After switching the potentials, the filament is pulse-heated to desorb collected fusion products followed by laser ionization. Before starting the next cycle and unblocking the primary beam, both laser shutters are closed and the potentials are switched back for the next accumulation on the filament.

To ensure an optimum duty cycle, the beam break in the cycle should be as short as possible. However, the exact timing of the potential switching in the cycle is crucial to prevent any direct transport of incoming, positively charged fusion products onto the detector creating a background count rate independent of any laser interaction. Therefore, a delay of 0.3 s between beam blocking and the potential switching is considered in the cycle. An additional delay of 0.7 s between the switch of the potentials and pulse-heating the filament ensures a completed change to the guiding mode before the filament has reached the desorption temperature towards the end of the heat pulse. A typical RADRIS cycle for resonance ionization of incoming 155Yb fusion products (t1/2 =1.8 s) is shown in Figure 1b. For this isotope, the cycle features an accumulation on the filament for the accelerator beam-on period of 3 s and a beam break of 3 s.

#### **3. New Detector Developments**

The total RADRIS efficiency, which corresponds to the ratio of detected laser ions and incoming fusion products, does not only depend on the transport efficiency and the detection efficiency, but also on the duty cycle due to the cyclic application of the RADRIS technique. In the current setup, the detection efficiency is limited to 40% due to the covered solid angle by the detector for alpha decay on the foil [13]. For short-lived nuclides and respective short cycles, additional losses due to radioactive decay reduce the overall efficiency and hamper the study of nuclides with half-lives t1/2 < 1 s. For long-lived nuclides on the other hand, longer beam breaks after a completed measurement point are necessary, reducing the duty cycle and efficient beamtime usage. Thus, the RADRIS cycle needs to be adapted for each isotope depending on its half-life [18]. New developments were initiated to improve the overall efficiency for the application of RADRIS to different nuclides.

#### *3.1. Rotatable Detector Setup*

In the previously used design of the RADRIS setup [13], each measurement point required a waiting time long enough to detect subsequent decays of resonantly ionized products collected on the detector. To allow for a more efficient usage of beamtime when longer-lived (>1 h) nuclides are studied, a rotatable multi-detector setup was developed. This setup enables to parallelize multiple measurements on long-lived nuclides by decoupling the collection phase on the detector from the measurements of subsequent alpha decays. The new design combines three identical PIPS detectors with an active area of 450 mm2 on a rotatable feedthrough, as shown in the schematic drawing in Figure 2a. The three detectors are positioned such that one of these detectors is placed on-axis with the transport electrodes to collect generated laser ions (the collection mode), while the other detectors are positioned off-axis to register residual alpha activity on their surfaces (the detection mode). After the collection is concluded on the on-axis detector, the detector

setup is rotated such that the next detector is in the collection mode, now for laser ions produced by light of the excitation laser tuned to the next wavelength.

In this new setup, recoil ions are directly guided onto the detectors without an additional collection foil in front. In this way, the detection efficiency for each detector can be increased from 40% to 50%. The increased tailing in the alpha spectra due to the decay of collected nuclides on the detector surface itself does not impact alpha signals from decays that differ by more than 0.25 MeV, as can be seen in the spectrum in Figure 2a.

**Figure 2.** (**a**) Schematic drawing of the rotatable RADRIS detector setup. Laser ions (blue) are guided towards the PIPS detector in on-axis position in the collection mode. For the next collection, the detectors are rotated such that the next detector is placed on-axis with the chamber. Subsequent decays on the prior collection detector are registered in an off-axis detection position. An alpha spectrum on one of the detectors is shown for accumulated 254Fm RIS signals of a measurement time of approximately 5 days. (**b**) Schematic drawing of the movable double-detector setup. Laser ions (blue) are guided towards the on-axis PIPS detector in the collection mode. For the detection mode, the detector is moved on top of a second PIPS detector to detect alphas emitted in the hemisphere opposite the active area of the collection detector.

The new detector arrangement was recently commissioned and tested in a first on-line laser spectroscopy of 254Fm with a half-life of 3.2 h. These data are still under analysis and will be published independently. Figure 2a includes accumulated alpha spectra of the 254Fm laser ions on one of the three detectors. With only a single detector, approximately one day per scan step of the excitation laser was expected to be required, while using the rotatable detector setup shortened the measurement time per detector and wavelength step to only seven hours. The resulting time gain allowed performing this experiment in 5.5 days with parasitic beam (5 ms long beam pulses with a repetition rate of 5 Hz) from the UNILAC accelerator at the GSI.

#### *3.2. Movable Detector Setup*

New challenges for the RADRIS technique arise when applying it to heavier and more exotic nuclei for which an enhancement in the overall efficiency can be of utmost importance. One of these challenges is the search for atomic levels in lawrencium, as for instance the production yield of 255Lr (≈437 nb) is approximately one order of magnitude lower than for 254No (≈2050 nb) [19,20] which was previously investigated for the search for atomic levels in nobelium [9]. In addition, predicted atomic levels accessible for laser spectroscopy range from 20,000 to 30,000 cm−<sup>1</sup> [21–25], which would require an extended measurement time to identify a first atomic level. Thus, a combination of an enhancement in the sensitivity and efficiency of the detection method will benefit the search for atomic energy levels. To enhance the applicability of the RADRIS technique, a new setup with a movable detector was designed as shown in Figure 2b. This new detector system includes a double-PIPS detector setup, where both detectors have an active area of 600 mm2. The collection PIPS detector is placed on rails, enabling the movement of this detector between an on-axis position for laser ion collection and an off-axis position located on top of the second detector, where the active areas of both detectors face each other. The swap between both positions occurs via a fast pneumatically-coupled linear feedthrough with a 50 mm stroke in less than 1 s. During accumulation on the filament for a next wavelength step, the collection detector

is moved to rest on top of the second PIPS. The efficiency of detecting the alpha decay of collected nuclei is hence increased, as usually approximately 50% of the alphas emitted in the hemisphere opposite the active detector area would be lost. In this constellation, the second detector with a distance of approximately 4 mm between both active areas increases the rate of detected events. More than 65% of the fraction of alpha events seen by the collection detector can now be detected with the additional detector, which would usually be missed.

For future experiments on longer-lived species, the movable and rotatable detector designs can be combined by adding detectors opposite to those detectors in the detection mode on the rotatable setup.

#### **4. Short RADRIS Cycle Development**

With typical RADRIS cycles, laser spectroscopic investigations on short-lived nuclides (t1/2 < 1 s) are not feasible due to radioactive decay losses. Therefore, a new RADRIS cycle scheme for such nuclides was developed and applied to 154Yb in recent on-line measurements. This isotope with a half-life of only t1/2 = 0.4 s and a relatively high production rate compared to the heavy actinides represents an ideal test case for the application of RADRIS to short-lived nuclei. This nuclide is produced in complete fusionevaporation reactions of a 48Ca beam at a beam energy of 4.55 MeV u−<sup>1</sup> on a 112Sn target in the 112Sn(48Ca,6n)154Yb reaction channel at the SHIP separator. Figure 3a shows the alpha-decay spectrum of produced nuclides in this reaction after direct transport to a single PIPS detector. In addition to 154Yb, other isotopes such as 155,156Yb, and other radionuclides are present in the alpha spectrum as shown in Figure 3b, as different evaporation channels exist for the de-excitation of the compound nucleus 160Yb\*. In addition, decay daughters with significant alpha-branching ratios along the decay chains are present. As these nuclides from different evaporation channels have similar velocities, the velocity filter SHIP transports them to the focal plane. Thus, measured alpha-decay signals need to be carefully gated according to the respective alpha-decay energies for unambiguous identification.

**Figure 3.** (**a**) Alpha spectrum of indirectly and directly produced nuclei in the 48Ca+112Sn fusionevaporation reaction guided directly onto the PIPS detector after stopping inside of the gas cell. The signal strengths mostly reflect the respective production cross sections. Additionally, decay losses during ion transport through the gas cell impact the signal strength especially of the short-lived 154Yb. Data for this spectrum were collected for 50 min with parasitic beam of 5 ms long beam pulses with a repetition rate of 5 Hz from the UNILAC accelerator at the GSI. (**b**) Observed evaporation channels of the 48Ca+112Sn fusion-evaporation reaction for a beam energy of 4.55 MeV u<sup>−</sup>1. Shown are products and the subsequent daughter nuclides decaying via alpha decay which are mainly observed in the experiment.

To systematically understand the limitations of the cycles in terms of ion mobility in the gas cell, the transport time of 154Yb through the cell to the detector was investigated. Therefore, reaction products were created during 5 ms long pulses of the UNILAC accelerator with a repetition rate of 1 Hz and entered the RADRIS gas cell, where they were stopped and transported directly to the detector. Figure 4a shows the time evolution of energy-gated and time-binned alpha events of 154Yb as a function of the cycle time. Due to the limited mobility [6,26] of the Yb ions in the argon buffer gas, the ions reach the detector with a certain delay. The stopping distribution of the recoils in the gas cell [27] in addition to diffusion processes lead to an increased width of the distribution of the ions' arrival time at the detector. To determine the time required for the ions to reach the detector, one has to solve the differential equation describing the increasing number of 154Yb ions on the detector after every accelerator beam pulse with respect to their successive decay as

$$\frac{dN}{dt} = A \cdot e^{\frac{-\left(t - t\_0\right)^2}{\left(2 \cdot w^2\right)}} - \lambda \cdot N\_{\prime\prime}$$

and fit the obtained solution for *N*(*t*) to the measured distribution in Figure 4a. Here, *N* describes the number of ions, *A* the amplitude of the distribution, *t* the cycle time, *tc* the center of the time distribution, *w* its width, and *λ* = *ln*(2) *<sup>t</sup>*1/2 the decay constant of 154Yb. For this model, a Gaussian time distribution for the ions reaching the detector was assumed. The transport time, defined as the time required for 84.13% of the ions in the arrival-time distribution to reach the detector (corresponding to the centroid *tc* of the distribution plus 1*w*), was determined to be 0.33 s in argon buffer gas at a pressure of 90 mbar and a potential gradient of approximately 29 V/cm from the filament to ground potential on the detector.

**Figure 4.** (**a**) Time structure of decay signals from incoming nuclei produced by 5 ms pulses from the accelerator with a repetition rate of 1 Hz. The incoming ions were directly transported to the detector. The transport time marked in red is determined as the time needed for 84.13% of the ions within the assumed Gaussian time distribution (black) to arrive on the detector. For more details, see text. (**b**) New, short RADRIS cycle for resonance ionization spectroscopy of 154Yb. The overall cycle, as well as the beam break and the delay between beam stop and potential switch were generously shortened. The delay time between stopping of the accelerator beam and arrival of laser ions is marked in the cycle and needs to be considered for further decay losses.

Taking this boundary condition into account, a short RADRIS cycle as shown in Figure 4b was implemented. A simplified, faster configuration was tested in which only the filament potential is switched instead of switching the potential of multiple transport electrodes. To speed up the filament potential switching, a fast high-voltage switch (Behlke GHTS) was used, enabling a switching time of a few ms. With this rapid switching, the new, short RADRIS cycle features a waiting time between the potential switch and the

pulse-heating of the filament reduced to 0.2 s to allow settling down of the stopped recoil ions, thus reducing further decay losses. With this modification and the new potential configuration for ion accumulation on the filament, no additional losses in the ion transport were observed.

To determine the RADRIS efficiency with the short cycle, the short-lived 154Yb was resonantly laser-ionized together with the neighboring, longer-lived 155Yb. Hereby, the ratio of 154Yb/155Yb was measured to be *R* = 7.40(16)% in the focal plane [28] as can be seen also in the alpha spectrum in Figure 3a. For laser spectroscopy of both isotopes, a two-step RIS scheme as shown in Figure 5b was employed. A grating dye laser was deployed in broadband configuration with approximately 6 GHz linewidth for the first excitation step (FES). The second excitation step (SES) was provided by a high-power excimer laser. By discriminating the registered alpha-decay energies, laser-induced signals stemming from the two investigated Yb isotopes could be individually identified.

**Figure 5.** (**a**) Resonance ionization signal of 155Yb (black) and 154Yb ( blue). Both rates are normalized to their respective count rates for a direct transport of produced recoil nuclei per accumulated charge integral of the primary beam. (**b**) RIS scheme for laser spectroscopy of 154,155Yb. The first excitation step was provided by a dye laser, the second, non-resonant step for ionization by a high-power excimer laser.

#### **5. Results with the Short RADRIS Cycle Implementation**

For characterization of the short cycle, results from laser spectroscopy of the shortlived 154Yb isotope were compared with those obtained for 155Yb, providing information on the RIS efficiency for ytterbium isotopes of different half-lives. The energy-gated alphadecay signal in dependence of the wavenumber of the FES is shown in Figure 5a, and the applied RIS scheme is presented in Figure 5b. The RIS count rates of 154Yb and 155Yb were normalized to the primary-beam integral for comparison with the respective rate of guiding ions directly to the detector. With the short cycle, it was possible to detect 154Yb from resonant laser ionization for the first time. The limited resolution in the gas cell [27] did not allow resolving the hyperfine structure of three expected hyperfine components in 155Yb with a nuclear spin of *I* = 7/2. In addition, the expected isotope shift of around 1 GHz [29–32] is much smaller compared to the spectral linewidth of the FES laser of approximately 6 GHz and has therefore not been properly determined with the available statistics. Due to the large linewidth, the expected hyperfine splitting has not been observed to additionally contribute to the broadening of the resonance.

To determine the respective RIS efficiency, a Gaussian fit was applied to both isotope spectra to extract the maximum normalized count rate on resonance. From the ratio of the obtained RIS signal rate to the signal rate from direct transport to the detector, a halflife-dependent efficiency was determined, which is depicted in Figure 6 as black symbols. In addition to 154,155Yb, also the nuclides 254No (*t*1/2 = 51.2 s) and 252No (*t*1/2 = 2.46 s) were investigated with respect to RIS and direct transport count rates on the detector and

respective efficiencies added to the results in Figure 6. It has to be noted that for 252No and 254No different cycles, optimized for the respective half-lives, were applied.

To conclude if the experimentally determined efficiency dependence is fully governed by the expected decay losses and to exclude additional losses due to other effects, the total efficiency Cycle of the different cycles was calculated analogously to [18]. Decay losses during accumulation on the filament and the overall usage of beamtime were considered by using the equation

$$\epsilon\_{\rm Cycle} = \frac{t\_{1/2}}{\left(t\_{\rm beam} + t\_{\rm break}\right) \cdot \ln(2)} \left[1 - e^{\frac{-\ln(2)}{t\_{1/2}} \cdot t\_{\rm beam}}\right] \cdot e^{\frac{-\ln(2)}{t\_{1/2}} \cdot t\_{\rm delay}}.$$

Here, *t*1/2 is the half-life of the considered nuclide, *t*beam (2 s in the case of the short cycle) is the beam-on time which equals the accumulation time on the filament, *t*break (1.5 s for the short cycle) is the time of the beam break. For the calculation, a continuous production during the accumulation phase was assumed, followed by a decay of the collected population with the respective half-life. Additional losses during the beam break are considered by the last exponential function with *t*delay = 0.93 s being the time between the beam shutoff and the ions reaching the detector in their required transport time after the desorption and resonant ionization. The calculated efficiencies for the respective cycles and nuclide half-lives are shown as red symbols in Figure 6. As the trend is of most importance in the comparison of experimental with calculated values, both values for 254No were chosen to coincide in the graph. From the observed trend it becomes clear that the experimental behavior is fully described by decay losses. This enables forecasting the RADRIS performance for other exotic cases such as the isotope 251No, for which a RIS-to-direct transport ratio of 0.053 was calculated for the new short cycle of 3.5 s duration with a break of 1.5 s shown in Figure 4b.

**Figure 6.** Comparison of efficiencies for rates of laser ions (with the laser frequency tuned on resonance) relative to direct transport rates. Experimentally determined efficiencies are shown in black. Red data points show estimated efficiencies of the RADRIS cycles in relation to the experimental RIS efficiency of 254No. Overall efficiencies of the RADRIS setup are added with the right scale considering the known efficiency for 254No [9].

From previous experiments on 254No and 252No, the total efficiency of the setup is known as number of detected laser ions in resonance relative to the number of respective recoil ions in the focal plane. The overall efficiencies were previously determined to 6.4% ± 1.0% and 3.3% ± 1.0% for 254No and 252No, respectively [9]. Comparing the expectation values for 251No to 254No, a total efficiency of 1.1% is expected for the performance of the setup with the short cycle. For a future RADRIS experiment on the nobelium isotope 251No, with a production cross section of 30 nb [19], a RIS signal rate of approximately 2.5 ions per hour can be expected.

#### **6. Conclusions and Outlook**

Different aspects in improving the performance of the RADRIS technique were investigated in this work. A new rotatable detector assembly was successfully commissioned, extending the half-life range of nuclides accessible by RADRIS to half-lives of at least 3 h. With the newly implemented short RADRIS cycle, short-lived nuclides with half-lives of less than 1 s can now be studied, which was demonstrated on the short-lived isotope 154Yb with *t*1/2 = 0.4 s. From comparison with calculations, the expected efficiency for the application of this new cycle to 251No is sufficient to allow the first optical spectroscopy of this isotope. Further upcoming experiments will focus on the search for atomic levels in lawrencium, for which the newly developed movable double-detector design features an efficiency gain with the second detector giving a decisive benefit. For future experiments, a re-designed transport electrode structure in combination with the double-detector setup will soon be commissioned in the upcoming beamtimes at the GSI to further boost the overall efficiency of the RADRIS technique.

**Author Contributions:** Conceptualization, J.W., B.A., M.B., P.C., A.C., R.F., F.G., O.K., T.K., E.K., M.L., J.L., A.M., D.M., S.N., S.R., E.R.-H., E.R., J.R., E.R.-R., M.V., P.V.D. and T.W.; methodology, J.W., M.B., P.C., T.K., M.L. and S.R.; investigation, J.W., B.A., M.B., P.C., R.F., F.G., O.K., E.K., M.L., J.L., D.M., S.N., S.R., E.R-H., E.R. and E.R.-R.; software, P.C.; formal analysis, J.W. and S.R.; data curation, J.W. and S.R.; writing—original draft preparation, J.W.; writing—review and editing, M.B. and S.R.; visualization, J.W.; supervision, M.B. and S.R.; project administration, M.B. and S.R.; funding acquisition, M.B., P.V.D. and T.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work has been supported by the Bundesministerium für Bildung und Forschung (BMBF, Germany) under Project No. 05P18UMCIA. This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 861198–LISA–H2020-MSCA-ITN-2019. E.K., E.R.-R. and M.L. acknowledge funding from the European Research Council (ERC) under the European Union's Horizon 2020 Research and Innovation Programme (Grant Agreement No. 819957). P.V.D., A.C., R.F. and J.R. acknowledge funding from the Research Foundation – Flanders (FWO) and from the EOS (nr. 30468642) project of the FWO and F.R.S.-FNRS under the Excellence of Science (EOS) programme. T.W. acknowledges funding from the Bundesministerium für Bildung und Forschung (BMBF, Germany) under grant number 05P21RDFN1.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**

The following abbreviations are used in this manuscript:


#### **References**


## *Article* **Resolution Characterizations of JetRIS in Mainz Using 164Dy**

**Danny Münzberg 1,2,3,\*, Michael Block 1,2,3, Arno Claessens 4, Rafael Ferrer 4, Mustapha Laatiaoui 3, Jeremy Lantis 3, Steven Nothhelfer 1,2,3, Sebastian Raeder 1,2 and Piet Van Duppen <sup>4</sup>**


**Abstract:** Laser spectroscopic studies of elements in the heavy actinide and transactinide region help understand the nuclear ground state properties of these heavy systems. Pioneering experiments at GSI, Darmstadt identified the first atomic transitions in the element nobelium. For the purpose of determining nuclear properties in nobelium isotopes with higher precision, a new apparatus for highresolution laser spectroscopy in a gas-jet called JetRIS is under development. To determine the spectral resolution and the homogeneity of the gas-jet, the laser-induced fluorescence of 164Dy atoms seeded in the jet was studied. Different hypersonic nozzles were investigated for their performance in spectral resolution and efficiency. Under optimal conditions, a spectral linewidth of about 200–250 MHz full width at half maximum and a Mach number of about 7 was achieved, which was evaluated in context of the density profile of the atoms in the gas-jet.

**Keywords:** JetRIS; fluorescence spectroscopy; gas-jet; de Laval nozzle; nobelium

#### **1. Introduction**

The measurement of atomic transitions via laser spectroscopy is a versatile method for determining fundamental nuclear and atomic properties [1–4]. At the GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany, laser spectroscopy is used at the Separator for Heavy Ion reaction Products (SHIP) [5,6] with a focus on the heavy actinide and transactinide region [4,7,8]. The low production rates and short half-lives of these nuclides pose difficult experimental challenges and require highly sensitive techniques. Recent laser spectroscopic measurements were conducted successfully at GSI on nobelium isotopes produced through fusion-evaporation reactions at SHIP using the Radiation Detected Resonance Ionization Spectroscopy (RADRIS) technique [7,9], where reaction products are thermalized in an argon filled gas cell and collected on a tantalum filament. The ions are neutralized by collection on a metallic filament, which is subsequently heated to produce an atomic vapor for resonance ionization spectroscopy (RIS). Due to the pressure and temperature conditions in the gas cell, the spectral resolution is limited to about 3 GHz. This is often insufficient to resolve all individual hyperfine components of the studied optical transition, as, e.g., in the case of 253No [10]. Additionally, species with half-lives of less than approximately one second are inaccessible to the RADRIS technique due to decay losses during recoil ion collection. To overcome both of these limitations, JetRIS has been constructed for high-resolution resonance ionization spectroscopy in a hypersonic gas-jet [10]. JetRIS combines the high resolution of the in-gas-jet laser spectroscopy technique developed at KU Leuven [11–13] with the sensitivity of the ion collection and neutral desorption from a heated filament used in the RADRIS technique [14,15]. In the new approach presented here, after neutralization, the atoms are carried through a hypersonic nozzle to form a

**Citation:** Münzberg, D.; Block, M.; Claessens, A.; Ferrer, R.; Laatiaoui, M.; Lantis, J.; Nothhelfer, S.; Raeder, S.; Van Duppen, P. Resolution Characterizations of JetRIS in Mainz Using 164Dy. *Atoms* **2022**, *10*, 57. https://doi.org/10.3390/ atoms10020057

Academic Editor: Jean-Christophe Pain

Received: 29 April 2022 Accepted: 25 May 2022 Published: 28 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

low-temperature and low-density gas-jet, reducing the Doppler and collisional broadening effects and thus increasing the spectral resolution by an order of magnitude. The ability to transport neutral species makes it possible to run the system in a continuous mode instead of in cycles as is happening in RADRIS. The negative potential on the filament can be applied at all times in addition to heating, minimizing the before-mentioned decay losses. JetRIS is designed to achieve a spectral resolution of at least 400 MHz for the heaviest elements, allowing for a more precise determination of the nuclear moments. To understand the performance of different nozzles, we present here the characteristics of these in terms of Mach number, spectral resolution and homogeneity of the produced jet.

#### **2. Experimental Procedure**

#### *2.1. A Technical Overview of JetRIS*

JetRIS consists of a high-pressure gas cell (stagnation pressure *P*<sup>0</sup> of 80–125 mbar argon) used to stop and thermalize recoil ions from fusion-evaporation reactions after separation from the primary beam by SHIP and a lower pressure jet cell (background pressure *<sup>P</sup>* of 5 × <sup>10</sup>−<sup>3</sup> mbar–2 × <sup>10</sup>−<sup>2</sup> mbar), which is used for laser spectroscopy. Inside the gas cell, the thermalized ions are transported via an electric field created by a set of cylindrical electrodes toward a filament located at the front of the nozzle, as sketched in Figure 1. This filament, typically made of tantalum, is resistively heated, allowing for neutralization and desorption of atoms, which are subsequently transported by a gas flow into the jet cell through the de Laval nozzle, forming a well-collimated hypersonic gas-jet. This gas-jet features a low temperature and a low pressure, thus reducing the spectral linewidth while the collimation of the gas-jet is crucial to maintain the highest efficiency. Two laser beams are used in a cross-beam geometry to interact with the gas-jet, performing two-step resonance ionization spectroscopy. The laser for the first excitation step is propagating anticollinearly relative to the gas-jet, while the second step proceeds in a perpendicular configuration. While the perpendicular configuration reduces the power density of the laser light, it helps in avoiding ionization in the gas cell. The photo-ions are then guided around a 90◦ curve via a radio frequency quadrupole (RFQ) to a detector cell, where a channel electron multiplier (CEM) or silicon detector is located. A more detailed description of JetRIS can be found in [10]. In this technique, the nozzle determines the achievable resolution and the total efficiency from the collimation. Therefore, a thorough characterization is essential in understanding the performance of the setup. At KU Leuven, such nozzles are studied in detail using Laser Induced Fluorescence (LIF) in Cu I using pulsed laser radiation as well as the RIS of neutral Cu atoms [16,17].

In this study, we follow a different path by using LIF of neutral 164Dy, which is illuminated by light from a cw-diode laser, propagating anticollinearly to the gas-jet as sketched in Figure 2. With this technique, three different de Laval type nozzles were investigated, and they were designed for different operation pressures and differences in their contour. All of them feature a throat diameter of 1 mm. The first nozzle is intended for usage at low stagnation pressures of *P*<sup>0</sup> = 80 mbar and a background pressure of *<sup>P</sup>* = 2.5 × <sup>10</sup>−<sup>2</sup> mbar. The diverging part of this nozzle has a length of about 1 cm as sketched in Figure 3. From fluid dynamic calculations, a jet of approximately Mach 8 was expected. The second nozzle is optimized for high stagnation pressures around *P*<sup>0</sup> = 300 mbar, and here, the diverging part has a length of about 3 cm. This nozzle is identical to the nozzles investigated recently at KU Leuven [16]. The third nozzle, referred to as the mid-range nozzle, has a conic contour. Here, the diverging part has a length of 2 cm. This nozzle was a prototype for operation in an intermediate pressure range while being simple to machine. No simulations were performed to optimize the design of this nozzle, and its optimal operating conditions were not previously known. To seed the atoms into the gas-jet for these tests, the tantalum filament in front of the nozzle was replaced by a tantalum strip that was previously loaded with a sample and resistively heated until a suitable fluorescence signal was observed, but the temperature of the filament was not

measured. It can only be approximated from its color when glowing, with an estimated temperature of 1200 ◦C.

**Figure 1.** Schematic overview of JetRIS. On the left side are the cage and funnel electrodes. In front of the nozzle, there is a tantalum filament. After being evaporated from the filament, the atoms follow the flow of the buffer gas through the nozzle into the gas-jet. Here, two laser beams used in a cross-beam geometry resonantly ionizes the formerly neutralized species of interest. The ions are guided around a curve via a 90◦ bend RFQ and collected on an α-Detector.

#### *2.2. Fluorescence Characterization*

During the experiments presented in this work, we used one-step laser excitation in contrast to the two-step resonant ionization that will be used in online experiments. As no ions were produced, the RFQ structure was removed (*cf*. Figure 2) and the fluorescence of the seeded atoms provided a way to determine the density and homogeneity along the gas-jet. The atom source was installed next to the nozzle entrance in the gas cell, consisting of a folded piece of tantalum foil, which contained a piece of a few mg of 164Dy with an isotopic purity of about 95%. The usage of an isotopically enriched sample ensured the investigation of a single atomic line as only minor contributions to the fluorescence signal from other isotopes are present and the even-even isotope features no hyperfine structure splitting. The foil was resistively heated with an electric power of 15 W to produce a dysprosium vapor, which was carried to the nozzle by the gas flow. A self-built laser with a 405 nm laser diode (Thorlabs L405P20) in an external cavity in Litrow configuration with approximately 12 mW of laser power and a sub-megahertz linewidth was used to excite the 4f106s2 → 4f106s6p transition in Dy I at a wavelength of 404.5 nm and with a transition strength of 1.92 × 108 <sup>s</sup>−<sup>1</sup> [18]. The laser beam was expanded to form a circle of about 10 mm in diameter and was aligned to propagate anticollinear to the central axis of the gasjet. A Complementary Metal-Oxide Semiconductor (CMOS) camera (Zelux® CS 165 MU) with a quantum efficiency of 50% was used to capture the fluorescence light originating from the atomic deexcitation. A bandpass filter featuring about 40% transmission at 405 nm

and a bandwith of 10 nm was installed in front of the camera. Pictures of the fluorescence, as shown in Figure 4, were taken as a function of gas pressure, wavelength and exposure time, which did not exceed 26 s due to limitations of the software for the camera.

**Figure 2.** Schematic overview of JetRIS. On the left side are the cage and funnel electrodes, which are necessary for an online experiment but were not in use for the fluorescence measurements. In front of the nozzle, there is a tantalum filament that contains a piece of 164Dy foil. After being evaporated from the filament, the atoms follow the flow of the buffer gas through the nozzle into the gas-jet. Here, a cw-diode laser beam at approximately 405 nm wavelength resonantly excites the dysprosium, and the resulting fluorescence is captured using a CMOS camera. The camera was mounted at a 45◦ angle relative to the field of view of this schematic.

**Figure 3.** Cross sectional profiles of the characterized nozzles. The base and the diameter of the hole are identical for every nozzle.

**Figure 4.** Example picture of the fluorescence acquired with the CMOS camera. Shown is the midrange nozzle at the centroid frequency of the transition using a stagnation pressure of 100 mbar and a background pressure of 6.47 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar. The visible stripes in the jet are a property of the laser diode used. The red box indicates the region that was considered in the analysis.

The fluorescence intensity was averaged in the radial plane, normal to the flow direction, in order to obtain information of the performance characteristics along the jet. Due to averaging, the stripes visible in Figure 4 did not disrupt the analysis.

#### *2.3. Characterization of the Gas-Jet*

The recorded fluorescence intensity was used to determine the density of atoms in the gas-jet, as well as to study the effective spectral broadening and, thus, the temperature of the jet, while exciting the atoms around the resonance frequency, i.e., performing spectroscopy. The intensity was evaluated pixelwise along the length of the gas-jet with the intensity averaged across the jet for each pixel in *x* direction. For each pixel, the normalized fluorescence intensity was plotted as a function of the laser frequency. A Gaussian fit to the data provided the centroid frequency and the spectral linewidth of the resonance. A number typically used to describe a gas-jet is the Mach number *M*, which is defined as the quotient of the stream velocity and the local speed of sound. It gives us an easy-to-compare variable that convolutes the speed of the jet and the temperature. The Mach number *M* is calculated with the following [11].

$$M = \sqrt{\frac{2}{\gamma - 1} \left( \frac{T\_0}{T} - 1 \right)}.\tag{1}$$

Here, *γ* is the ratio of the specific heat capacities of the gas, which is 5/3 for a monoatomic gas, *T* is the temperature of the jet and *T*<sup>0</sup> is the initial temperature of the gas before it reaches the nozzle.

The temperature *T* of the jet was determined from the measured linewidth by using the following relation.

$$
\Delta \nu\_{\rm D} = 2 \sqrt{\ln(2)} \frac{\nu\_{01}}{c} \sqrt{\frac{2kT}{m}}.\tag{2}
$$

Here, Δ*ν*<sup>D</sup> is the contribution to the full width at half maximum (FWHM) of the Doppler broadening, *ν*<sup>01</sup> is the transition frequency, *c* is the speed of light, *k* is the Boltzmann constant and *m* is the mass of 164Dy. As the measured resonance features a Voigt profile, the Doppler broadening can be determined by using the following approximate relation.

$$
\Delta\nu = 0.5346 \,\Delta\nu\_{\rm L} + \sqrt{0.2166 \,\Delta\nu\_{\rm L}^2 + \Delta\nu\_{\rm D}^2}. \tag{3}
$$

Here, Δ*ν* is the FWHM of the measured Voigt profile, and Δ*ν*<sup>L</sup> describes the Lorentzian part of the overall resolution, which contains the natural linewidth and the pressure broadening. Any contribution from power broadening is neglected.

To obtain an estimate of the temperature *T*0, the connection between *T*<sup>0</sup> and the stream velocity was used [11].

$$
\mu = \sqrt{\frac{\gamma k T\_0 M^2}{m \left(1 + \left(\frac{\gamma - 1}{2}\right) M^2\right)}}.\tag{4}
$$

where *m* is the mass of the buffer gas, and *u* is the stream velocity of the jet, which can be obtained from the recorded centroid using the optical Doppler shift from the literature value for the transition of *ν*<sup>01</sup> = 24,708.97 cm−<sup>1</sup> [19]. The stated reference did not mention the isotope of dysprosium for the recorded value. Therefore, the transition was measured with JetRIS by shining a laser beam perpendicular to the flow direction of the gas-jet, which yields a value free from Doppler shift. The measured value was in agreement with the literature value reported in [19].

Above Mach 5, the stream velocity reaches 95% of its maximum. Since the Mach number was expected to be around 5–8, the mean value of these (*M* = 6.5) was taken as an approximation for *T*0. The deviation of the temperatures obtained with *M* = 5 and *M* = 8 from the value at *M* = 6.5 is around 3% and was considered when determining the uncertainty of the experimentally determined Mach number. A typical value of *T*<sup>0</sup> obtained from the fitting of the data is 380 K, indicating some heating of the gas from the hot filament, a fact that was already observed in previous investigations in Leuven [16].

To determine the quality of the gas-jet, a new metric was established and will be referred to as the homogeneity factor *H*. The photon density of the fluorescence light was used to evaluate the homogeneity of the sample atom density across the full length of the jet and was compared to a hypothetical, perfectly homogenous jet of constant light intensity. For this new factor, two different integrals have been calculated. A normalized integral of the hypothetical perfect jet *I*max, where the intensity should be constant over the entire length of the jet and the intensity integral over the experimental intensity values *I*, resulting in the following equation.

$$H = \frac{\int I}{\int I\_{\text{max}}}.\tag{5}$$

The boundaries of both integrals are the same and are determined by the length of the real jet. The homogeneity factor provides a simple value between 0 and 1, where 0 would mean no observed fluorescence, meaning no formation of a jet, and 1 would mean that we would observe a perfectly homogenous jet.

For an overall view on the performance of a nozzle, the spectral resolution and *M* were multiplied by the relative intensity, summed up and divided by the sum of the relative intensities, therefore making an intensity-weighted average. The uncertainty of these parameters was determined as the standard deviation of the individual numbers.

#### **3. Results**

#### *3.1. Resolution*

The low-stagnation-pressure nozzle is a de-Laval nozzle and was designed for a stagnation pressure of *<sup>P</sup>*<sup>0</sup> = 80 mbar and a background pressure of *<sup>P</sup>* = 2.5 × <sup>10</sup>−<sup>2</sup> mbar [20]. However, while investigating the resolution as a function of the background pressure, as shown in Figure 5a for the low stagnation pressure nozzle, a lower background pressure was found to provide the best resolution. This general trend that the resolution is improving as the background pressure drops was observed for all three nozzles. It has to be noted that the available pressure ranges are limited by the capacity of the JetRIS pumping system. At optimal parameters, the best achievable resolution was 212 ± 4 MHz for the lowstagnation-pressure nozzle, 239 ± 13 MHz for the mid-range nozzle and 311 ± 15 MHz for the high-stagnation-pressure nozzle. These values are all intensity-weighted averages of the individual values for each pixel slice along the gas-jet. It was verified whether analyzing the jet as a whole has an impact on the values relative to the pixel-by-pixel analysis, and

both methods are in agreement with one another. The optimal stagnation pressure of 300 mbar for the high-stagnation-pressure nozzle could not be reached, again limited by the pumping system [16]. Under the available conditions, the low-stagnation-pressure and mid-range nozzle outperformed the high-stagnation-pressure nozzle with regards to the obtained resolution. All of the obtained spectral linewidths are smaller than the stated goal of 400 MHz [10]. The individual parameters for the measurements are summarized in Table 1.

**Figure 5.** (**a**) Intensity-weighted averages of the resolution for the low-stagnation-pressure nozzle. The resolution generally improves as the background pressure is reduced. (**b**) Resolution along the jet for different parameters of the low-stagnation-pressure nozzle. (**c**) Example Gaussian fit of the intensity as a function of the laser frequency for the low-stagnation-pressure nozzle at *P*<sup>0</sup> = 80.6 mbar and *<sup>P</sup>* = 5.6 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar. In both pictures, the parameters are as follows: black: *<sup>P</sup>*<sup>0</sup> = 80.6 mbar and *<sup>P</sup>* = 5.6 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar; red: *<sup>P</sup>*<sup>0</sup> = 80.6 mbar and *<sup>P</sup>* = 7.0 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar; green: *<sup>P</sup>*<sup>0</sup> = 80.0 mbar and *<sup>P</sup>* = 8.5 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar; blue: *<sup>P</sup>*<sup>0</sup> = 82.2 mbar and *<sup>P</sup>* = 10.4 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar.



#### *3.2. Mach Number*

The jet was evaluated for its Mach number as described in Section 2.3. For the determination of the gas-jet temperature from the linewidth, the spectral profiles were fitted with Gaussian profiles. The calculated linewidths were taken as the values for the overall resolution, since the fit was in good agreement with the data, as shown in Figure 5c. The natural linewidth can be calculated to be 30.5 MHz from the transition strength. The pressure broadening could not be determined; however, in the case of Cu I studied at Leuven, it was found to be approximately 3 MHz [16]. Power broadening has been neglected due to the low laser power used in the experiment. The two factors were added to a Lorentzian contribution to the linewidth of 33.5 MHz and the temperaturedependent part of the resolution was calculated according to Equation (3). With this, the Mach numbers were calculated as *M* = 7.2 ± 1.0 for the low-stagnation-pressure nozzle, *M* = 7.2 ± 1.0 for the mid-range nozzle and *M* = 4.6 ± 0.5 for the high-stagnation-pressure nozzle in the best case, respectively. The Mach numbers of the high stagnation pressure nozzle are significantly lower than *M* = 8, expected from fluid-dynamics calculations and from observations at KU Leuven [16]. This was most likely due to the fact that the nozzle was used outside of its desired pressure range [21]. In this investigation, some small uncertainties in the evaluation remain, concerning the determination of the stagnation temperature *T*<sup>0</sup> and the frequency instability of the laser diode while measuring, which are expected to be reflected in the uncertainties. The presented values agree very well with the observations from previous studies at KU Leuven, albeit it has to be noted that the measurements were taken under different conditions. On one hand, the investigations in Leuven for the low-stagnation-pressure nozzle were performed using only the central 1 mm diameter of the gas-jet core with collinear illumination, while in this work the entire jet is illuminated anticollinearly, adding the jet boundary layer of the jet in the evaluation. Furthermore, the 164Dy atoms used in this study are heavier than the 65Cu atoms used in Leuven, which is a much lighter system that is closer to the carrier gas (40Ar).

#### *3.3. Homogeneity Factor*

Finally, the homogeneity as defined in Equation (5) was evaluated, which shows a quite different behavior of the nozzles. With the best possible parameters, a value of *H* = 0.33 was achieved for the low-stagnation-pressure nozzle, compared to values of *H* = 0.63 for the mid-range nozzle and *H* = 0.76 for the high-stagnation-pressure nozzle. The corresponding intensity distributions along the jet for the three nozzles are shown in Figure 6.

Clearly, none of the investigated nozzles provide an ideal jet with a perfectly homogenous density profile and some losses from diffusion into the background gas are unavoidable. Furthermore, the intensity profile at the spectral maximum was compared with the intensity profile averaged over the spectral profile. The latter corresponds to the total density of the jet independent of the velocity distribution and shows a better homogeneity. Nevertheless, the profile at the maximum excitation frequency corresponds to the accessible fraction of the density and, thus, provides a better estimate on the expected efficiency. It shall be noted that the pulsed laser for the intended resonant ionization application features a significantly larger bandwidth of about 100 MHz compared to the sub-megahertz bandwidth of the cw diode laser used in this work [22]. This will enable addressing more atoms of the ensemble in the gas-jet and, thus, the effective homogeneity will be in between the two curves in the upper and lower panels of Figure 6, respectively.

The homogeneity factor provides a good impression about the achievable efficiency from the atom density along the gas-jet, but it does not yet provide conclusive information about the overall efficiency of JetRIS. Further measurements are planned, including the transport efficiency of atoms evaporated from the filament and transported through the nozzle to the detector at online-like conditions. For this, a radioactive recoil source will be used, since it releases ions at a known rate, allowing for a quantitative measurement, independent from ionization efficiency when using lasers.

**Figure 6.** Intensity distribution at the centroid frequency (**upper**) and average over all frequencies (**lower**) for the best parameters of the low-stagnation-pressure nozzle (black, *P*<sup>0</sup> = 80.6 mbar, *<sup>P</sup>* = 5.6 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar), mid-range nozzle (red, *<sup>P</sup>*<sup>0</sup> = 125 mbar, *<sup>P</sup>* = 7.35 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar) and highstagnation-pressure nozzle (green, *<sup>P</sup>*<sup>0</sup> = 125 mbar, *<sup>P</sup>* = 9.3 <sup>×</sup> <sup>10</sup>−<sup>3</sup> mbar).

#### **4. Summary and Outlook**

To enable high-resolution laser spectroscopy of the heaviest elements at GSI, Darmstadt JetRIS is under development. For an online experiment, efficiency is of paramount importance while maintaining a high spectral resolution. Therefore, fluorescence spectroscopy was performed to characterize three hypersonic nozzles in terms of spectral resolution, Mach number and homogeneity. These nozzles were designed for operation at different stagnation pressures. For each nozzle, gas pressures were identified resulting in a resolution sufficient for determining the hyperfine structure of 253No, for example. The highest spectral resolution was found for the low-stagnation-pressure and the midrange nozzle with linewidths of 211 MHz and 239 MHz, respectively, for the investigated ground-state transition at 404.5 nm in 164Dy. In contrast, the high-stagnation pressure nozzle provided a linewidth of 335 MHz at the intended operation pressures of up to 125 mbar. The larger mass of 253No compared to 164Dy should allow achieving a higher resolution, but since the transition used for nobelium has a wavelength of 333 nm [7] compared to 405 nm for dysprosium, the resolution can be expected to be the similar in both cases. The low- and mid-range nozzle show a similar performance in terms of the Mach number as well. In terms of jet homogeneity, the high-stagnation-pressure nozzle showed the best performance. The mid-range nozzle seems to be the best overall choice, since its resolution and homogeneity are both close to the optimal values found for the other two nozzles. According to investigations at KU Leuven, the high-stagnation pressure nozzle would greatly benefit from operation at a higher stagnation pressure [16]. Nevertheless, our obtained resolution is already close to the value of 170 MHz projected in [16] for laser spectroscopy in the actinide region. Further studies will be performed offline with radioactive sources and resonance ionization spectroscopy to determine the efficiency of JetRIS before measuring online isotopes of nobelium at GSI, Darmstadt.

**Author Contributions:** R.F. and S.R. conceived the experiment. S.R., D.M., J.L. and S.N. set up the diode laser system. D.M., J.L. and S.N. conducted the measurements. D.M. and J.L. analyzed the data with the input from P.V.D., R.F., M.L., S.R., M.B. and A.C. The paper was written by D.M. with the input from P.V.D., M.B., A.C., R.F., J.L., M.L., S.N. and S.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** M.L. acknowledges funding from the European Research Council (ERC) under the European Union's Horizon 2020 Research and Innovation Programme (Grant Agreement No. 819957).

**Data Availability Statement:** The data presented in this study are available upon request from the corresponding author.

**Acknowledgments:** The authors thankfully acknowledge the LARISSA group (Institut für Physik, Johannes Gutenberg-Universität Mainz) for the contribution of the laser diode cavity, as well as the target lab of GSI, Darmstadt, for the supply of enriched 164Dy.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **A Progress Report on Laser Resonance Chromatography**

**Elisa Romero Romero 1,2,3,\*, Michael Block 1,2,3, Biswajit Jana 1,2,3, Eunkang Kim 1,2,3, Steven Nothhelfer 1,2,3, Sebastian Raeder 2,3, Harry Ramanantoanina 1,2,3, Elisabeth Rickert 1,2,3, Jonas Schneider 1, Philipp Sikora <sup>1</sup> and Mustapha Laatiaoui 1,2,3**


**Abstract:** Research on superheavy elements enables probing the limits of nuclear existence and provides a fertile ground to advance our understanding of the atom's structure. However, experimental access to these atomic species is very challenging and often requires the development of new technologies and experimental techniques optimized for the study of a single atomic species. The Laser Resonance Chromatography (LRC) technique was recently conceived to enable atomic structure investigations in the region of the superheavy elements. Here, we give an update on the experimental progress and simulation results.

**Keywords:** laser spectroscopy; superheavy elements; laser resonance chromatography

#### **1. Introduction**

In the last two decades, there have been outstanding and exceptional efforts in the discovery and study of the superheavy elements [1]. One of the highlights is the completion of the seventh row in the periodic table with the addition of four new synthetic elements in 2016, including oganesson (Og, element number *Z* = 118), the last and heaviest element to date. The development of new selective and efficient techniques has had an impact on the discovery of these elements and their detailed study. Some of these elements are predicted to not behave chemically like their lighter homologs, with relativistic effects being the dominant cause of this peculiarity [2,3].

The challenges to study them are manifold. Superheavy elements are produced in nuclear fusion-evaporation reactions using powerful accelerators at extremely low rates in the presence of a huge background from primary-beam particles. In addition, they usually exist only for a few seconds after their production, which explains why their basic chemical and atomic properties are often not known [1]. Efficient gas chromatography has been used to elucidate the adsorption enthalpies. The heaviest element studied with this technique is flerovium (Fl, *Z* = 114), with a half-life ranging between one and two seconds [4–6]. A few years ago, experiments using surface-ionization techniques were successfully applied to lawrencium (Lr, *Z* = 103), aiming at establishing the element's ionization potential [7].

Deeper insights into the atomic properties and structure can be gained from optical spectroscopy. At present, in-gas-cell laser resonance ionization spectroscopy [8–10] is the most advanced method for atomic structure studies on the heaviest elements. A recent breakthrough in this research field was achieved with the spectroscopy of nobelium (No, *Z* = 102) [10] using the RAdiation-Detected-Resonance-Ionization-Spectroscopy (RADRIS) technique.

Our alternative way of optical spectroscopy, namely, Laser Resonance Chromatography, has already been proposed for optical spectroscopy of lawrencium ions and is explained in detail in Ref. [11]. Briefly, this technique combines resonant laser excitation with electronic-state chromatography [12–15] and is conducted directly on the ion in-situ,

**Citation:** Romero Romero, E.; Block, M.; Jana, B.; Kim, E.; Nothhelfer, S.; Raeder, S.; Ramanantoanina, H.; Rickert, E.; Schneider, J.; Sikora, P.; et al. A Progress Report on Laser Resonance Chromatography. *Atoms* **2022**, *10*, 87. https://doi.org/ 10.3390/atoms10030087

Academic Editor: Christian Parigger

Received: 2 August 2022 Accepted: 31 August 2022 Published: 6 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

without the need for a neutralization step. Given the fusion products are stopped and extracted from a gas catcher in a +1 charge state, a laser of a proper wavelength optically pumps the ions into a metastable state from the ionic ground state. After this step, the ions are injected into a drift tube filled with diluted helium (He) gas, where they undergo a constant drift under the influence of an external electric field. Different interactions of the ions in the different states with helium result in state-specific ion mobilities, which enable modern electronic-state chromatography, i.e., separating the ions in the ground state from the metastable ions by drift time [13,16]. In other words, the changes in the arrival time distributions caused by laser excitations give the resonance signal. Although only applicable to ions and dependent on the presence of metastable states, the method of electronic state chromatography is well established for many elemental cations of the first-, second- and third-row transition metals [17–22].

To this end, a first-generation drift tube chamber has been designed for LRC applications. This design is different from traditional ion-mobility-experiment applications since suppressing deactivation of metastable states is mandatory and is pursued by reducing the length of the tube and operating at relatively low pressures.

In the next two sections, we give a brief report on the experimental progress of the laser resonance chromatography project by presenting the experimental apparatus including the laser system, the cryogenic drift tube, and the corresponding ion trajectory simulations. Due to the scarcity of data and since scandium (Sc, *Z* = 21) can be deemed as a homolog of lutetium (Lu, *Z* = 71), the later simulations were conducted for singly charged scandium in its ground and metastable states of known ion mobilities. In the last section, we give prospects of LRC experiments on Sc+, Lu<sup>+</sup> and its heavier iso-electronic system, Lr+. A summary of the important properties of these elements is compiled in Table 1.

**Table 1.** Relevant electronic states in Sc+, Lu+, and Lr<sup>+</sup> ions. The experimental ion mobilities are given for a helium temperature of 295 K. Predictions are marked by .


*a*: Refs. [23–25]; *b*: Refs. [26–28]; *c*: Ref. [28–30]; *d*: Ref. [31].

#### **2. Experimental Approach**

Laser resonance chromatography couples laser spectroscopy with ion mobility spectrometry. It is based on a population transfer between metastable ionic states in a resonant laser excitation process. A laser excites the ion, e.g., from the ground state to an intermediate level (to be optically probed) in an allowed optical transition. The intermediate level depopulates partly to lower-lying metastable states that do not easily quench to the ground state. In a simplified picture, the ion changes its size during this process, which can then be exploited for purposes of diagnostics. The resonant process is identified by a change in the characteristic arrival-time distributions of the ions on a particle detector after passing a drift tube filled with helium gas at pressures < 10 mbar. Since the mobility is function of the gas temperature and could be distinct for the different states, the operation of the drift tube at cryogenic temperatures usually provides an additional degree of freedom to optimize time resolution and state separation [19,32].

The method is generally applicable for transition-metal ions including Lu<sup>+</sup> and Lr+. Promising optical pumping schemes for singly charged rutherfordium, the next-heavier element within the fourth row transition metals, have already been proposed [33]. Compared to many existing spectroscopy techniques, the LRC approach has a number of key advantages, some of which are:


#### *2.1. The LRC Apparatus*

The LRC setup is shown in Figure 1. It consists of five different pressure sections (PS) for stopping, extraction, separation, mass selection and detection of the sample ions. The ionized residual nuclei produced during the fusion-evaporation process lose most of their kinetic energy by passing through a metallic window of a few μm thickness before they are thermalized by collisions with the He buffer gas inside of the stopping cell (PS1) at a pressure of about 60 mbar. The thermalized ions are ejected through a convergent– divergent nozzle of 0.6 mm throat diameter towards a radio frequency quadrupole (RFQ) in PS2 that serves to extract and further cool the ions and to guide them towards the buncher in the next pumping section (PS3). The first-generation stopping cell of the SHIPTRAP setup together with its extraction RFQ i used for this purpose [34].

**Figure 1.** Schematic overview of the LRC apparatus. See text for more information.

The subsequent buncher installation enables a spatial confinement of the ions for laser spectroscopy and a precise referencing of their arrival time distributions. It consists of four stainless steel rods with diameters of 3.5 mm, each divided into 25 segments. The distance between opposite rods is 2*r*<sup>0</sup> = 3 mm. PS3 also incorporates a cryogenic drift tube and an ion guide. The drift tube is used for electronic-state chromatography and is explained in more detail in Section 2.3. The ion guide comprises 10 segments of similar geometry as the buncher segments and is used to transport and focus the ions into the quadrupole mass filter (Extrel QMS) in the pumping section PS4, where the ions are selected based on their mass-to-charge ratio. Next, the ions are focused by einzel lenses and a X and Y steerer towards the detection system, which contains a channeltron detector (Dr Sjuts K15) installed in the last pumping section PS5.

#### *2.2. The Laser System*

One of the characteristics and a potential advantage of the LRC technique compared with conventional resonance ionization spectroscopy is the use of only one laser beam to search for optical resonances by optical pumping of metastable states. The laser system is shown in Figure 2. It consists of a 10-kHz Nd:YAG laser (Edgewave, 90 W at 532 nm) that pumps a dye laser (Sirah Credo), providing laser pulse energies between 10 and 100 μJ in the ultraviolet (UV) range from 220 up to 360 nm after frequency doubling or tripling, depending on the dye. Both lasers are installed next to the LRC apparatus and the laser beam path and optics are arranged as shown in Figure 2. The fundamental wavelength is monitored using a wavelength meter (HighFinesse WS7 UVU) featuring autocalibration via an integrated calibration source. For initial experiments and offline studies, a Nd:YAG laser (Continuum Minilite II) operated at 10 Hz repetition rate was used in addition to produce ions via ablation from primed samples inserted inside of the stopping cell (PS1).

**Figure 2.** LRC laser system. Edgewave Nd:YAG laser pumps a Sirah Credo dye laser. Laser ablation is carried out using a Continuum Minilite II Nd:YAG laser. The wavemeter is used for wavelength monitoring. Abbreviations: M, mirror; L, lens; TP, telescopic lens; C, cylindrical lens; BS, beam splitter; BW, Brewster window; G, grating; PE, prism expander; DC Res, dye cell resonator; DC Amp, dye cell amplifier; OC, output coupler; FCU, frequency conversion unit.

#### *2.3. The Drift Tube Outer Chamber*

The main components of the drift tube section are shown in Figure 1 (PS3). The section incorporates two stainless steel chambers: The outer vacuum chamber and the cryogenic drift tube are connected to the buncher on the left side and to the ion guide on its exit on the right side. The outer chamber has a cuboid shape with edge lengths of (L × W × H) = (255 mm × 269 mm × 262 mm). A 1600 l/s turbomolecular pump (TMP, Edwards STP 1603C) is connected to this chamber via a DN-200 ConFlat flange to pump it down to pressures <10−<sup>8</sup> mbar in standby mode or <10−<sup>2</sup> mbar in operation mode. The chamber provides vacuum and thermal shielding for the cryogenic drift tube and features high voltage and RF feedthroughs, a gas inlet, pressure gauges, electrical feedthroughs for heaters and temperature sensors, view ports for the laser beam and a DN-63 ConFlat flange to connect a free piston Stirling cryocooler (CryoTel-CT). The latter has a cooling capacity of about 11 W at 77 K and is connected to the drift tube via four copper strands with a cross sectional area of 16 mm2.

#### *2.4. The Drift Tube Inner Chamber*

The cryogenic drift tube sits at the heart of the LRC apparatus. A schematic overview of this is shown in Figure 3. The tube has a hexagonal shape with an inner diameter of 46 mm and a length of 53.5 mm. It is fixed to the outer chamber via 12 titanium spokes (M2, DIN 975/DIN 976 Titanium Grade 2). The drift tube chamber is plated on the outside with a thin copper layer of 50–100 μm thickness to enable better heat conductance and a homogeneous distribution of the temperature over the whole drift tube during cooling and warming phases. It includes a gas inlet and outlet, a connection for a pressure gauge (Pfeiffer Vacuum PKR 360) and several tapered holes to fix heaters (high power resistors TCP100U) and temperature sensors (Lake Shore Germanium-CD). The tube has octagonal flanges at both ends that also serve to attach it to the spokes on the outer chamber.

In its interior, the drift tube chamber incorporates eight stainless steel electrodes of 20 mm inner diameter, 24 mm outer diameter and of a width of 5 mm. Six of the electrodes are enclosed by two identical end caps designed to have an electrode in one side and a diaphragm of 1 mm on the other side, serving either as injection or exit nozzle. The caps also serve as a support for the stainless steel fixation of the buncher and the ion guide; cf. Figure 3. All inner electrodes are electrically connected to each other by seven 1-MΩ resistors in series to build up a resistance of 7.15 MΩ between the end caps. The electrodes are supported via Vitronit ceramic rods of 5 mm diameter and 46.5 mm length and separated by 0.5 mm from each other with Vitronit ceramic cylindrical spacers of 5 mm length and 5 mm inner diameter. All inner electrodes plus the end caps are surrounded by a ceramic cylinder of 40 mm inner diameter , 44 mm outer diameter and length of 46.5 mm to isolate the electrodes from the grounded tube housing.

**Figure 3.** 3D cross-sectional view of the LRC drift tube and its components.

#### **3. Ion Drift Simulations**

Simulations were performed for the drift tube using the SIMION software package [35] in order to estimate Sc<sup>+</sup> drift times at a given He pressure and temperature and to extract suitable voltage configurations to be used in future LRC experiments. In the simulations, we considered both Statistical Diffusion (SDS) and Viscous Damping (VD) models [36]. Hard sphere model simulations could not be performed thus far due to the lack of reference data for the collision cross sections.

The reduced mobility for Sc<sup>+</sup> in the ground state and Sc+<sup>∗</sup> in the metastable state for the SDS and VD models were taken from [20]; cf. Table 1. We tested two configurations, one with a roughly constant electric field (unfocused beam) and a second with a gradually increasing electric field (focused beam). Ions were generated at the entrance of the drift

tube and their drift times were recorded when they exited through a 1 mm or 2 mm diameter nozzle to explore the feasibility of enhancing the transmission while keeping time resolution unchanged. The two voltage configurations we used are shown in Table 2. We simulated different voltage values of U0 in a way that the resulting average ratio of electric-field strength to gas number density, E/n0, spanned a range between 1 and 30 Td, with 1 Td = 10−<sup>17</sup> <sup>V</sup> · cm2. For each value of U*o*, 10, 000 ions were generated in a 3D Gaussian distribution with a standard deviation of *σx*,*y*,*<sup>z</sup>* = 0.2 mm. The ion mobilities were calculated from the reduced mobilities by considering a pressure of 2 mbar of helium gas at a temperature of *T* = 297 K. Figure 4a shows ion trajectories projected on the symmetry plane obtained for the two different electric field configurations at an average reduced field of 15 Td. Using the VD model, the electric fields acting on each ion during its drift were recorded for each voltage configuration and allowed us to extract the mean electric field; cf. Figure 4b.

**Table 2.** Voltage configurations for unfocused (\*) and focused beam (\*\*). n is the electrode number. In the case of the unfocused beam, the voltages applied to the different electrodes were scaled by n, whereby values of U0 between 0.1–10 V were applied in steps of 0.1 V to span a range of E/n0 between 1–30 Td. For the focused beam, we added different offsets to the unfocused beam configurations as given in the table, where *δ* = 0.75 V and U0 was varied between 0.1–2.7 V to span in total an E/n0 range between 1–30 Td.


**Figure 4.** (**a**) Trajectories for E/n0 = 15 Td for unfocused (\*) and focused beams (\*\*). (**b**) Electric fields along the electrodes for focused (**top**) and unfocused (**bottom**) beams.

#### *Results*

We made a comparison of the transmission efficiency between focused and unfocused beam configurations and between 1 mm and 2 mm exit nozzles for Sc<sup>+</sup> ions. To this end, we defined this efficiency as the fraction of the number of ions arriving at the exit nozzle within a radius of 0.5 mm (1 mm) from the center axis respective to the initial number of ions for 1 mm (2 mm) nozzle diameters. Figure 5 shows this transmission efficiency for the different states as function of the reduced field. For the unfocused beam using the 1 mm diameter exit nozzle, it grows quickly as the reduced field increases to reach a maximum of 3.7% at an E/n0 value of about 8.5 Td and decreases with increasing fields to stagnate at about 2%. When using the 2 mm diameter exit nozzle, it grows even quicker as the reduced field increases to reach a maximum of 4.6% at an E/n0 value of about 7.6 Td and stays stable up to around 20 Td to then increase again. According to the simulations, there is nearly no difference for the unfocused beam in terms of transmission efficiency between the ground and the metastable state for both 1 mm and 2 mm exit diameters below 20 Td. In the case of a focused beam using the 1 mm exit nozzle, the efficiency follows the unfocused beam up to 3 Td, then gradually increases with increasing fields, but stays below that achieved for the unfocused beam till reaching 15 Td. In the case of the 2 mm nozzle, the focused beam and the unfocused beam have similar transmission up to 10 Td. From these values onward for both nozzles, 1 mm and 2 mm, the deviation in the transmission of the two states becomes apparent. A maximum transmission efficiency of 7% for 1 mm and 20% for 2 mm nozzles in the focused beam configuration is achieved for the ground state ions at 30 Td, while the metastable state transmission stagnates at about 5% and 15% for 1 mm and 2 mm, respectively. Since the ions in the ground state exhibit a higher mobility, they can drift faster compared with the ions in the metastable states and thus are less prone to transversal diffusion losses. Theoretically, even higher efficiencies can be expected for these latter scenarios if the reduced field is increased beyond 30 Td, but only at the cost of deactivating states due to gas collisions that would degrade the metastable signal [17,37,38]. In addition, increased electric fields carry the risk of gas discharges with only a few hundred volts for 2 mbar of He gas and lead to shorter drift times due to higher velocity, which can in turn lead to both neutralization and a lower resolving power, respectively.

**Figure 5.** Ion transmission for unfocused (\*) and focused (\*\*) beams for both ground (blue) and metastable states (red) and for (**a**) 1 mm and (**b**) 2 mm exit diameter nozzles. Dash-line: focused beam; solid line: unfocused beam.

To better understand the behavior of the different electric field configurations in terms of resolution, we analyzed the drift time differences between the ground state and metastable state. Figure 6 shows the drift time for the two Sc<sup>+</sup> states in the different beam configurations in the case of the 1 mm nozzle; similar behavior was observed for the 2 mm exit diameter nozzle. It becomes apparent that, irrespective of the electronic states, the gradual increase of the electric field (corresponding to the beam focusing scenario) cause the ions to drift at small velocities the majority of the time and to lag behind in comparison when

they are exposed to an average but rather homogeneous electric field (unfocused beam). The relative drift time differences exhibit a maximum at reduced field values between 5 Td and 10 Td in both configurations, indicating the best time resolution. However, a deeper insight is obtained by including peak broadening effects in the analysis by comparing the time histograms of the transmitted ions from the simulations. The different ionic states can be disentangled better from each other at larger reduced fields, which means at smaller absolute drift times and thus at the cost of relative drift time differences. If we compare the two different configurations, it becomes clear that, here as well, the unfocused beam provides better working conditions because it provides better time resolution over a larger range of E/n0 values. In the case of a focused beam, the time peaks can be partly disentangled only at fields higher than ≈10 Td. The 2 mm configuration shows a similar trend with respect to the resolution. However, even though the larger nozzle provides higher efficiency, this latter is not so remarkably higher as to trade off the vacuum conditions inside of the PS3 section. We can consider this option for future tests.

**Figure 6.** Absolute drift time comparison between ground and metastable states for both focused and unfocused beams. Unfocused beam: = ground state;  = metastable state. Focused beam: = ground state; = metastable state. Insets: selected histograms for E/n0 = 5, 15, and 30 Td to demonstrate resolution behavior. \* = unfocused beam ; \*\* = focused beam; blue = ground state; red = metastable state.

#### **4. Current Status and Outlook**

In the summer of 2022, the LRC setup is nearly complete and the commissioning phase has already begun with testing of the vacuum and functionality of key components such as the buffer gas stopping cell, the quadrupole mass filter and the laser systems. The cryogenic drift tube together with the miniature ion guide and buncher are being assembled and are ready for integration into the setup. Different ion sources are available, including a laser ablation source and a 223Ra recoil ion source, with the latter being best suited for optimizing and quantifying the transmission efficiency through the whole apparatus.

SIMION simulations were performed for the LRC drift tube using two electric field configurations: unfocused and focused beams; and two geometry configurations: 1 mm and 2 mm exit nozzle diameters. From these simulations, we inferred that a rather homogeneous electric field enables a comparably higher ion transmission while maintaining a good resolution at relatively low E/n0 values using a 1 mm exit diameter nozzle. In addition, it can be expected that working at lower fields minimizes the risk of gas discharges and deactivation of states. This can therefore also be very beneficial for a successful application of the LRC method. Higher transmission can be achieved when using a 2 mm exit nozzle and eventually focusing the beam into the nozzle.

Our first proof-of-principle experiments will target 45Sc<sup>+</sup>, cf. Table 1. These offline measurements are currently being prepared and we expect them to last for up to one year. The relative mobility difference between the ground and the metastable state is about 20% at 295 K in He [25], which should be sufficient to enable LRC measurements. The laser

probing occurs inside the buncher, i.e. before the ion drift, via laser resonant excitation of the *z*3D◦ <sup>1</sup> state at 27,917.78 cm−<sup>1</sup> to optically pump the ion into the metastable state *<sup>a</sup>*3F2 at 4802.87 cm−1. Since the metastable state is energetically relatively close to the ground state, deactivation of states will likely occur during the ion-atom collisions [39]. If such collisional de-excitations dominate and entirely prevent the chromatography of Sc+, we will pursue LRC experiments on 175Lu<sup>+</sup>, the lighter chemical homologue of Lr+. In the Lu<sup>+</sup> experiments, we will probe the 3P◦ <sup>1</sup> state at 28,503.16 cm−<sup>1</sup> that feeds the 3D1 metastable state at 11,796.24 cm<sup>−</sup>1. Since this latter state is energetically high enough above the ground state, level crossings in the corresponding diabatic potential curves become unlikely. Thus, for short drift paths, as in the LRC experiments, we expect deactivation of states to be suppressed in Lu+-He collisions, particularly at moderate kinetic energies. Here, one should note that in Ref. [25], the signal of the metastable state could still be observed even for Sc<sup>+</sup> drifting inside a drift tube of about 2 m length. However, since the drift tube of the LRC apparatus is only 45 mm long, the chromatography will require detailed analysis of the arrival-time distributions due to expected moderate time resolution; cf. Figure 6.

Applying the LRC technique to stable Lu ions can give us a better understanding of the trade-off we should make to achieve maximum count rates without losing the chromatography information. Once experimentally optimized for low yields, LRC can then be applied to search for atomic levels in the heavier iso-electronic system 255Lr<sup>+</sup> in on-line experiments.

**Author Contributions:** Conceptualization, M.L.; Setup design, E.R.R and M.L.; Simulations, E.R.R.; Simulations Methodology, E.R.R. and M.L.; Formal analysis, E.R.R.; Funding acquisition, M.L.; Investigation, E.R.R. and M.L.; Project administration, M.L.; Resources, B.J., E.K., H.R., E.R. and J.S.; Software, E.R.R.; Supervision, M.B. and M.L.; Validation, E.R.R, M.B., S.R. and M.L.; Visualization, E.R.R.; Writing—original draft, E.R.R; Writing—review & editing, E.R.R., M.B., M.L., S.N., S.R. and P.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 Research and Innovation Programme (Grant Agreement No. 819957).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We would like to thank the workshops of the Chemistry, Nuclear Chemistry, Physics and Nuclear Physics departments of the JGU Mainz.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**

The following abbreviations are used in this manuscript:


#### **References**


## *Article* **Extending Our Knowledge about the 229Th Nuclear Isomer**

**Benedict Seiferle \*, Daniel Moritz, Kevin Scharl, Shiqian Ding †, Florian Zacherl, Lilli Löbell and Peter G. Thirolf \***

Faculty of Physics, Ludwig-Maximilians University München, Am Coulombwall 1, 85748 Garching, Germany; daniel.moritz@physik.uni-muenchen.de (D.M.); K.Scharl@physik.uni-muenchen.de (K.S.); dingshq@gmail.com (S.D.); Zacherl.Florian@physik.uni-muenchen.de (F.Z.); Lilli.Loebell@physik.uni-muenchen.de (L.L.)


**Abstract:** The first nuclear excited state in 229Th possesses the lowest excitation energy of all currently known nuclear levels. The energy difference between the ground- and first-excited (isomeric) state (denoted with 229mTh) amounts only to <sup>≈</sup>8.2 eV (≈151.2 nm), which results in several interesting consequences: Since the excitation energy is in the same energy range as the binding energy of valence electrons, the lifetime of 229mTh is strongly influenced by the electronic structure of the Th atom or ion. Furthermore, it is possible to potentially excite the isomeric state in 229Th with laser radiation, which led to the proposal of a nuclear clock that could be used to search for new physics beyond the standard model. In this article, we will focus on recent technical developments in our group that will help to better understand the decay mechanisms of 229mTh, focusing primarily on measuring the radiative lifetime of the isomeric state.

**Keywords:** Th-229; nuclear clock; hyperfine structure spectroscopy; ion trap

#### **1. Introduction**

The nuclear first excited state in 229Th is in the focus of nuclear as well as atomic physics research. Due to its low excitation energy in the range of ≈8.2 eV (we took the mean value of the two most recent energy determinations [1,2]), the first nuclear excited state plays an exceptional role with the possibility to be excited by laser light. This led to the proposal to use the 229Th nucleus as a basis for a nuclear optical clock [3]. It has been predicted that a nuclear clock could potentially reach a relative frequency uncertainty in the range of 10−<sup>19</sup> [4]. Therefore, such a nuclear clock could complement current atomic clocks. It could especially be employed in the search for new physics beyond the standard model [5].

The reader is referred to the references [6–8] for a detailed overview of the topic.

The exceptionally low excitation energy plays an important role when one considers the possible decay channels of the isomer: The isomer potentially decays to its ground state via four decay channels: *γ* decay, internal conversion (IC), bound internal conversion (BIC) [9] and electronic bridge (EB) [10]. In the gamma decay channel, the isomer decays by emitting a photon that carries the excitation energy. The partial lifetime of this decay channel has been predicted to be in the range of 103 to 104 s [11,12].

The *γ*-decay channel competes with the internal conversion decay channel, whose lifetime has been measured in neutral atoms to be in the range of several microseconds [13], making it orders of magnitude faster than the *γ* decay. During the internal conversion decay, the energy of the isomeric state is transferred to the electronic shell and an electron is emitted into the vacuum. A prerequisite for the IC decay to occur is that the binding energy of one of the bound electrons (which is given by the ionization potential) is below the isomeric excitation energy. For the specific case of 229mTh, the IC decay is already energetically forbidden for 229mTh1<sup>+</sup> ions with an ionization potential of <sup>≈</sup>12 eV.

**Citation:** Seiferle, B.; Moritz, D.; Scharl, K.; Ding, S.; Zacherl, F.; Löbell, L.; Thirolf, P.G. Extending Our Knowledge about the 229Th Nuclear Isomer. *Atoms* **2022**, *10*, 24. https:// doi.org/10.3390/atoms10010024

Academic Editor: Camillo Mariani

Received: 31 December 2021 Accepted: 31 January 2022 Published: 14 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Another possible decay channel is bound internal conversion, where the decay energy is also transferred to the electronic shell. Instead of an electron being emitted, as in the internal conversion decay, an electronic state is excited. A requirement for electronic bridge decay is the presence of a transition in the electronic shell that is in resonance with the isomeric ground-state transition. This strong requirement is relaxed in the electronic bridge channel, where the isomer decays by exciting a virtual state in the electronic shell, which subsequently decays to a real electronic state. The excess energy is then carried away in the form of photons.

In the following, we focus on prospects to measure the radiative decay channel in a new setup currently being arranged at LMU Munich.

#### **2. Towards Radiative Lifetime Measurements**

For the measurement of the radiative lifetime it is envisaged to monitor the number of 229Th ions in the isomeric state 229mTh over time. The measurement of the radiative lifetime requires the complete suppression of all the other competing decay channels, such as internal conversion and bound internal conversion. The suppression of internal conversion can be achieved by preventing 229mTh ions from neutralizing, since IC is energetically forbidden in Th ions. Therefore, the ions are confined in an ion trap. The trap is operated under cryogenic conditions to achieve the high vacuum quality that is needed to realize long storage times in the range of the expected radiative lifetime.

The appearance of BIC and EB can be excluded by measuring the lifetime in different electronic states. For a successful measurement of the lifetime, the storage time of the 229(m)Th ions (the m in brackets indicates that we are dealing with a cloud of ions in the nuclear ground state and the nuclear isomeric state) in the ion trap needs to be at least in the range of the expected lifetime, i.e., several 1000 s. This requires optimum vacuum conditions that can only be achieved under cryogenic conditions at temperatures around 4 K.

The general concept of the setup is shown in Figure 1. 229(m)Th3<sup>+</sup> ions extracted from a buffer gas stopping cell are loaded axially (from the left side in Figure 1) into a cryogenic linear Paul trap. 229(m)Th3<sup>+</sup> ions are used due to their favorable electronic level scheme exhibiting a rather simple alkali-like structure of an inert Rn core and a single valence electron, providing a closed three-level Lambda system suitable for for laser excitation and fluorescence detection. There, they are sympathetically cooled by 88Sr ions, which are provided by an ion source and are axially loaded into the same linear Paul trap from the opposite side (i.e., the right side in Figure 1).

**Figure 1.** Visualization of the experimental setup and concept. A detailed explanation is provided in the text. 229(m)Th ions are produced in the *α*-decay of 233U and extracted by a buffer-gas stopping cell and an extraction radio frequency quadrupole (RFQ). A quadrupole mass separator (QMS) allows for the selection of a specific charge state, which enables the loading of 229(m)Th <sup>3</sup><sup>+</sup> into a cryogenic Paul trap. Sr ions produced by an ion dispenser source are bent by 90° by an electrostatic ion bender and are then injected into a QMS that selects 88Sr<sup>+</sup> from other naturally occurring Sr isotopes. 88Sr<sup>+</sup> ions are loaded into the cryogenic Paul trap and can be laser cooled.

The Sr ion source is placed 90° off axis, and the ions are bent by 90° using an electrostatic bending quadrupole in order to prevent the cryogenic stages from being exposed to the heated ion source and thereby reducing the heat load to the cold stages. This geometry also allows for a direct line of sight along the central axis of the setup (e.g., to align lasers along the axis for Doppler cooling and spectroscopy).

#### *2.1. Stopping Cell and Extraction RFQ*

229(m)Th ions are produced in the *α* decay of 233U, where the isomeric state is fed by a 2% decay branch. For this reason, a 233U *α* recoil source with an activity of 10 kBq is placed in a buffer-gas stopping cell. The source consists of a Si wafer disk with a diameter of 30 mm. 233U is deposited onto the disk by electroplating. 229(m)Th ions leaving the source material with a kinetic energy of ≈84 keV are stopped in 32 mbar catalytically purified helium. The ions are guided by an RF-DC funnel towards a de-Laval nozzle (nozzle diameter Ø = 0.4 mm) that connects the high-pressure stopping cell to another vacuum chamber.

The RF-DC funnel consists of concentrically stacked ring electrodes, whose inner diameter is reduced linearly with the distance from the source, thus, creating a funnel-like shape. A DC gradient along the funnel electrodes guides the ions axially towards the de-Laval nozzle. Sinusodial RF-fields that are varying in phase by 180° between neighboring funnel electrodes prevent the ions from hitting the electrodes. The electrical potentials together with the funnel-like geometry allow the transport of ions that are far from the central axis towards the nozzle exit.

In the vicinity of the de-Laval nozzle, a gas flow drags the ions through the nozzle and injects them into the subsequent chamber. The formed supersonic gas jet is generated by a pressure difference between the buffer-gas stopping cell (typically at 32 mbar) and the subsequent chamber, which is typically pumped to a pressure in the range of 10<sup>−</sup>3–10−<sup>4</sup> mbar.

This chamber houses an axially segmented radio-frequency quadrupole (RFQ). A voltage gradient along the axis drags the ions through the remaining buffer gas, while the applied RF voltage keeps the ions on the central axis. This enables the formation of a cooled ion beam.

#### *2.2. Quadrupole Mass Separators*

The setup contains two quadrupole mass separators (QMS 1 and 2). QMS 1 is used to generate an isotopically pure 229(m)Th3<sup>+</sup> ion beam that can be injected into the Paul trap and is located between the extraction RFQ and the cryogenic Paul trap. The second QMS (QMS 2) is placed between the cryogenic Paul trap and the 88Sr ion source. QMS 2 serves two purposes: First, it is used to select 88Sr ions from the ion beam generated by the Sr ion source.

In addition to other naturally occurring Sr isotopes, the ion beam may also contain elements other than Sr (such as K, Rb or Cs) due to the production process of the source. Secondly, QMS 2 can be used to investigate a possible formation of molecules of the Th ions after being trapped in the Paul trap. The QMS modules follow the design of [14], which was also used in earlier experiments [1,13,15]. In order to achieve the required mass resolving power, the RF-voltage amplitudes are actively stabilized by an FPGA-based circuit. For further details, see [16].

#### *2.3. Cryogenic Paul Trap*

The central structure of the setup is a Paul trap that is designed to be operated at cryogenic conditions. The design of the Paul trap follows closely the design used in [17,18]. For further details, see [16].

#### *2.4. Sr Ion Source and Ion Bender*

The Sr ion source is a commercially available heated dispenser ion source. The source is typically heated to a temperature above 1000 °C by applying a current of ≈2.2 A to a heating filament that is part of the source assembly.

The ions that are emitted from the source are extracted and focused by two ring electrodes. An electrostatic ion bender, consisting of four quarter cylinders that form a quadrupole potential, bends the ions by 90° towards QMS 2. Before entering QMS 2, the ions pass three more ring electrodes that help to efficiently inject them into QMS 2.

#### *2.5. Cooling Lasers and HFS Lasers*

To resolve the hyperfine-structure (HFS) shifts that are used to distinguish between the nuclear ground and nuclear isomeric state the 229(m)Th3<sup>+</sup> ions need to be cooled.

Direct laser cooling of 229(m)Th3<sup>+</sup> has already been achieved [19]. In our setup, 229(m)Th3<sup>+</sup> ions are sympathetically cooled by 88Sr ions, whose mass-to-charge ratio (88 u/e) is close to that of 229(m)Th3<sup>+</sup> (76.3 u/e). The lack of hyperfine-structure shifts in 88Sr ions provides a simpler cooling scheme than for 229(m)Th3<sup>+</sup> . Doppler cooling can be performed on the <sup>2</sup>*S*1/2 →<sup>2</sup> *<sup>P</sup>*1/2 transition at 422 nm [20]. The ions are re-pumped from the <sup>2</sup>*D*3/2 to the <sup>2</sup>*P*1/2 state with 1091 nm radiation (see the left part of Figure 2). The discrimination between the nuclear ground state and nuclear isomeric state is performed by measuring the HFS of 229(m)Th3<sup>+</sup> .

The HFS will be probed on the <sup>2</sup>*F*5/2 →<sup>2</sup> *<sup>D</sup>*5/2 transition at 690 nm. An additional re-pumping laser at 984 nm is needed to pump from the <sup>2</sup>*F*7/2 level back to <sup>2</sup>*D*5/2. The level scheme is shown in the right panel of Figure 2. 229(m)Th3<sup>+</sup> exhibits a rich hyperfine structure; therefore, in order to avoid pumping into (hyperfine) dark-states, corresponding sidebands are generated with electro-optic modulators (EOMs). All central wavelengths are provided by external cavity diode lasers.

The 422 nm laser is locked to a close-by transition in Rb and shifted with an acoustooptical modulator (AOM) by approximately 440 MHz in order to drive the transition in 88Sr [21]. The remaining lasers will be stabilized by either using a scanning transfer cavity or a commercial wavelength meter.

**Figure 2.** The relevant level schemes of singly charged 88Sr and triply charged 229Th . The presence of the hyperfine structure is indicated by the broadened width of the bars.

#### *2.6. Measurement Scheme*

The measurement scheme involves two stages. Ions are loaded into the trap and cooled down in a first stage. The second stage involves the measurement of the lifetime. First, 229(m)Th3<sup>+</sup> ions are loaded into the trap. The ions are extracted from a buffer gas stopping cell. We estimate the number of extracted ions by scaling the number of 229(m)Th3<sup>+</sup> ions extracted from a similar buffer-gas stopping cell and a similar source geometry [15] with the source activity. In Ref. [15], the number of extracted 229(m)Th3<sup>+</sup> ions was on the order of 10<sup>4</sup> ions per second with a source activity of 290 kBq. Therefore, we expect an extraction rate in the range of 102 ions per second.

This number, however, requires experimental verification, as the exact extraction rate is influenced by several factors, such as the buffer gas cleanliness. We expect a small number of 229(m)Th3<sup>+</sup> ions in the range between 10 and 100 to be loaded into the trap. It is envisaged to form an ion crystal by sympathetic cooling and to identify the nuclear state of the trapped ions by measuring their hyperfine structure. When there is at least one isomer confined in the trap, the lifetime measurement is started. This involves imaging the fluorescence radiation of individual ions onto an (EM)CCD camera.

This allows for identification of the decay of the isomer by tagging ions in the isomeric state on the camera image via their HFS fluorescence and registering their decay to the ground state; the lasers are set to exclusively drive HFS transitions that correspond to the isomeric state. When the isomeric state decays to the ground state, the respective thorium ion turns dark on the camera.

In order to double-check that the ion was not lost due to any other process (i.e., neutralization or molecule formation), the laser is set to drive nuclear ground-state HFS transitions immediately after the ion has turned dark. If the ion is still present in the trap, the time of the decay event can then be recorded and used for data analysis.

It is possible that the isomeric radiative lifetime is affected by the electronic state. For cross-checks, the duty-cycle of the 690 nm laser can be varied. This will leave the ions in the electronic ground-state for a variable amount of time. Additionally, by varying the duty cycle of the re-pumping laser (984 nm), it is possible to pump the ions into the <sup>2</sup>*F*7/2 electronic state and investigate the isomeric lifetime for ions in this electronic excited state.

#### **3. Conclusions**

We presented a setup that is able to measure the radiative lifetime of 229mTh in the absence of the internal conversion decay channel. For this purpose, triply charged 229(m)Th ions are confined in a cryogenic Paul trap. 229(m)Th is cooled sympathetically by a laser-cooled cloud of 88Sr ions. The number of 229mTh ions is monitored over time by measuring the hyperfine-structure shifts specific for 229mTh .

**Author Contributions:** The original draft was prepared by B.S. with input from B.S., D.M., K.S., S.D., F.Z., L.L. and P.G.T.; supervision, P.G.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work is part of the 'ThoriumNuclearClock' project that received funding from the European Research Council (ERC) under the European Union's Horizon 2020 Research and Innovation Programme (Grant Agreement No. 856415) and by the European Union's Horizon 2020 Research and Innovation Programme under grant agreement No 664732 "nuClock".

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We thank J. R. Crespo López-Urrutia, E. Peik, M. Okhapkin, J. Thielking, J. Weitenberg and L. v.d. Wense for fruitful discussions and their support.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **First Offline Results from the S3 Low-Energy Branch**

**Jekabs Romans 1,\*,†, Anjali Ajayakumar 2, Martial Authier 3, Frederic Boumard 4, Lucia Caceres 2, Jean-François Cam 4, Arno Claessens 1, Samuel Damoy 2, Pierre Delahaye 2, Philippe Desrues 4, Antoine Drouart 3, Patricia Duchesne 5, Rafael Ferrer 1, Xavier Fléchard 4, Serge Franchoo 5, Patrice Gangnant 2, Ruben P. de Groote 1, Sandro Kraemer 1, Nathalie Lecesne 2, Renan Leroy 2, Julien Lory 4, Franck Lutton 2, Vladimir Manea 5, Yvan Merrer 4, Iain Moore 6, Alejandro Ortiz-Cortes 2, Benoit Osmond 2, Julien Piot 2, Olivier Pochon 5, Blaise-Maël Retailleau 2, Hervé Savajols 2, Simon Sels 1, Emil Traykov 7, Juha Uusitalo 6, Christophe Vandamme 4, Marine Vandebrouck 3, Paul Van den Bergh 1, Piet Van Duppen 1, Matthias Verlinde 1, Elise Verstraelen 1and Klaus Wendt <sup>8</sup>**


**Abstract:** We present the first results obtained from the S3 Low-Energy Branch , the gas cell setup at SPIRAL2-GANIL, which will be installed behind the S<sup>3</sup> spectrometer for atomic and nuclear spectroscopy studies of exotic nuclei. The installation is currently being commissioned offline, with the aim to establish optimum conditions for the operation of the radio frequency quadrupole ion guides, mass separation and ion bunching, providing high-efficiency and low-energy spatial spread for the isotopes of interest. Transmission and mass-resolving power measurements are presented for the different components of the S3-LEB setup. In addition, a single-longitudinal-mode, injectionlocked, pumped pulsed-titanium–sapphire laser system has been recently implemented and is used for the first proof-of-principle measurements in an offline laser laboratory. Laser spectroscopy measurements of erbium, which is the commissioning case of the S3 spectrometer, are presented using the 4 *<sup>f</sup>* 126*s*2 3*H*<sup>6</sup> <sup>→</sup> <sup>4</sup> *<sup>f</sup>* <sup>12</sup>(3*H*)6*s*6*<sup>p</sup>* optical transition.

**Keywords:** resonance ionization laser spectroscopy; gas cell; hypersonic gas jets; radio frequency quadrupoles; nuclear ground state properties; isotope shift; hyperfine structure

**Citation:** Romans, J.; Ajayakumar, A.; Authier, M.; Boumard, F.; Caceres, L.; Cam, J.-F.; Claessens, A.; Damoy, S.; Delahaye, P.; Desrues, P.; et al. First Offline Results from the S3 Low-Energy Branch. *Atoms* **2022**, *10*, 21. https://doi.org/10.3390/ atoms10010021

Academic Editor: Alexander Kramida

Received: 7 January 2022 Accepted: 3 February 2022 Published: 9 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

The Super Separator Spectrometer (S3) [1] is a fusion–evaporation recoil separator, which is currently under construction at the SPIRAL2 facility in GANIL, aiming to study exotic neutron-deficient isotopes in the actinide and super-heavy element regions, and in the *N=Z* region around 100Sn [2]. The fusion–evaporation reactions will be produced by an intense heavy ion beam, impinging on a thin target. The low-production crosssections and the available primary beam intensities at various facilities worldwide limits the production rates, and thus the amount of experimental data of very exotic nuclear systems. To overcome this obstacle, the superconducting LINAC of the SPIRAL2 facility has been developed to produce stable ion beams from He to U with energies from 0.75 up to 14.5 MeV/u, and intensities from 1pμA up to Ni [1]. Primary beams of such high intensities will make SPIRAL2-S<sup>3</sup> and its low-energy branch (S3-LEB) a prominent place to study the ground and isomeric state properties of exotic nuclei [3]. For a detailed description of the SPIRAL2 project, one can refer to [4].

The S3-LEB will be installed at the S3 final focal plane for some of the first experimental campaigns, and it will deploy a variety of low-energy measurement techniques (laser spectroscopy, decay spectroscopy and mass spectrometry). The underpinning working principle of the S3-LEB setup is the in-gas laser ionization and spectroscopy (IGLIS) technique [5,6], which aims to perform laser spectroscopy measurements to extract the isotope shifts and hyperfine parameters of radioactive isotopes. This experimental data can give access to differences in mean square charge radii *<sup>δ</sup>r*2, magnetic dipole <sup>μ</sup> and electrical quadrupole *Q* moments, as well as nuclear spins *I*, which are crucial for validating atomic and nuclear models, and for improving our understanding of the atomic and nuclear structure in poorly explored regions of the nuclear chart. However, the access to *I* and *Q* can be highly case-dependent, due to line-broadening mechanisms. One such example is the predicted existence of the island of stability of super-heavy elements [7].

Together with the hot-cavity laser ion sources used at ISOL facilities [8,9], IGLIS belongs to the broader class of laser ion source and laser spectroscopy techniques which probe the radioisotopes very close to the production or stopping area. These techniques allow the production of element-selective ion beams with high efficiencies. Nevertheless, their spectral resolution is typically limited by broadening mechanisms. The hot-cavity spectroscopy is dominated by a large Doppler broadening, induced by the *T* ∼ 2000 ◦C temperature of the laser beam–atom interaction region. At ISOL facilities, it is thus common to study radioactive beams after reacceleration and mass separation using high-resolution collinear fluorescence [10,11] or resonance ionization spectroscopy (RIS) [12]. Recently, new approaches for improving the spectral resolution of hot-cavity laser spectroscopy have also been explored, with promising results (such as the use of perpendicular illumination [13] and Doppler-free, two-photon spectroscopy [14]).

With the IGLIS method, one first thermalizes and neutralizes the reaction products in the buffer gas of a gas cell that is kept under a constant gas flow. Performing laser ionization spectroscopy in such an environment results in spectral line widths of several GHz, due to collisional broadening. A crucial upgrade for the IGLIS technique has been the use of a de Laval nozzle at the exit of the gas cell, which creates a collimated and homogeneous hypersonic gas jet of low temperature *T* and low density *ρ* [6], containing the products of interest. Such an environment allows for laser spectroscopy with reduced broadening mechanisms by about an order of magnitude, while maintaining a high selectivity and efficiency [5].

The S3-LEB setup has been developed by a collaboration between KU Leuven, SPIRAL2-GANIL, LPC Caen, IJCLab, University of Jyväskylä and University of Mainz. The setup is currently being commissioned at the GANIL Ion Source using Electron Laser Excitation (GISELE) [15] and LPC Caen. In this paper, the S3-LEB setup will be described and some first results from the offline commissioning tests will be presented.

#### **2. The S**<sup>3</sup> **Low-Energy Branch**

*2.1. Gas Cell, RFQ Ion Guides and Mass Spectrometer*

The starting point of the S3-LEB setup is a gas cell, in which the S3 fusion–evaporation recoils will enter via a thin window. A 3D image of the gas cell is presented in Figure 1. The next point of the setup is the beam transport, mass separation, bunching and cooling stages. This is achieved by the static and alternating electric fields created by a set of radio frequency quadrupole (RFQ) structures. An image of the full RFQ chain is presented in Figure 2.

Once stopped in the buffer gas environment, neutralization and thermalization of recoils will occur by interactions with the gas atoms and the electron density created by the stopped ion beam. The gas cell follows closely the design currently used at KU Leuven [16]. It is designed to be operated with argon gas at 200–500 mbar under constant flow, which exits the cell through a de Laval nozzle, having typically a 1 mm throat diameter. Gas flow simulations using COMSOL [16,17] have been performed in order to optimize the gas cell geometry and find an optimal volume providing an efficient stopping and extraction of the S<sup>3</sup> beam, while maintaining minimal extraction time. The resulting internal cross-section of the gas cell has a 30 mm depth and a 70 mm width. With this geometry, simulations give an average extraction time from the stopping area to the exit hole of about 500 ms for a 1 mm throat diameter.

A feedthrough in the gas cell body allows the insertion of two filaments that are resistively heated for evaporating an element used in the offline tests or as an online reference. The gas cell body and filament holder flange are water-cooled and the entire gas cell can be baked by resistively heated cartridges inserted in the gas cell body. The temperature is monitored by PT100 sensors. Just before the exit of the gas cell, two ioncollector electrodes are installed for removing non-neutralized ions in online experiments.

**Figure 1.** 3D cross-sectional view of the S3-LEB gas cell.

The gas cell has three laser windows, two just before the exit, facing each other, and one opposite and concentric to the exit hole. At the gas cell exit, a de Laval nozzle is installed, the geometry of which is optimized using the calculations performed by the Von Karman Institute for Fluid Dynamics (VKI, Belgium) [18]. On the exit side of the gas cell, aligned with the nozzle, two extraction plates—one on ground potential and the other on a slightly positive potential—provide an initial guiding field for the ions towards the RFQ chain.

The RFQ design follows the initial concept from KU Leuven [19], with further adaptations. For each RFQ, the RF voltage is impedance-matched using a specially designed

transformer with a tunable capacitor connected to the secondary circuit. DC voltage gradients are applied via voltage divider resistor chains across the RFQs.

First, the ions enter a segmented *S*-shape RFQ (SRFQ), which has the purpose of extracting the ions from the jet and decoupling the laser and ion beam axes. The SRFQ is located in the same vacuum chamber as the gas cell; thus, it is in a relatively high-pressure environment for RFQ operation (∼10−2–10−<sup>1</sup> mbar). The SRFQ has two injection plates that can be biased, and a linear DC gradient is applied on top of the RF voltage, to drag the ions through it. At the end of the first straight section of the SRFQ, a mirror fixed on top of the structure guides the laser light longitudinally into the gas cell.

**Figure 2.** Full S3-LEB ion guide layout. From left to right: mobile ion source, SRFQ, mRFQ, QMS and RFQ*cb*. See text for details.

After the SRFQ, the ions enter the mini-RFQ (mRFQ), which serves as a differential pumping stage and hence stands between two areas of approximately two orders of magnitude different vacuum levels. The vacuum chambers of the SRFQ and of the mRFQ only communicate through a 3 mm-radius bore of the latter. The gas load in the SRFQ area is pumped by an Edwards GXS450 screw pump, while in the mRFQ area, a Pfeiffer STPiXA3306C turbo pump, coupled to an Edwards GXS160F screw pump for pre-vacuum, are used to remove the remaining gas.

Next, ions enter the quadrupole mass filter (QMF), which was designed to reach a mass-resolving power *m*/(2Δ*m*) of ∼50. The first and last QMF segments can be DC biased independently from the rest, allowing it to act as a Brubaker lens [20].

After the QMF, the ions enter the cooler–buncher RFQ (RFQ*cb*), which is a two-section system. In the cooler section, which is surrounded by a metallic housing, the ions are cooled by helium gas, which is injected at the center of the RFQ. This minimizes the longitudinal and transversal emittance of the beam. In the following buncher section, the ions are bunched using a potential well created by a series of segments that are connected to high-voltage switches. After a predefined trapping time in the buncher, the extraction takes place by switching the trapping voltages to an extraction ramp, which accelerates the ions out of the RFQ*cb*. Differential pumping stages separate the QMF from the poor intermediate pressure areas of both the mRFQ and the cooler.

Once the cooled and bunched ion beam leaves the RFQ chain, it enters the pulse up (PU) drift tube. The tube is used for ion beam reacceleration up to ∼3–3.5 keV kinetic energy, which is the design voltage for the final point of the S3-LEB setup, consisting of a multi-reflection, time-of-flight (MR-TOF) mass spectrometer. When the ions enter the PU drift tube, its electrode is biased at ∼−1.5 kV. When the ions are at its center (typical flight times from the buncher are between 5 and 10 μs), the electrode voltage is switched to ∼+1.5 kV. This gives the ions a relative kinetic energy gain of ∼3 keV.

Further beam purification and detection will be performed by the MR-TOF mass spectrometer, called Piège à Ions Linéaire du GANIL pour la Résolution des Isobares et la mesure de Masse (PILGRIM). In this device, the ion beam is reflected between two electrostatic mirrors until it is separated in time of flight, leading to a mass-resolving power *<sup>R</sup>* = *<sup>m</sup>*/(2Δ*m*) ≈ <sup>10</sup><sup>5</sup> [21]. The setup will expand in its capabilities by a decay spectroscopy setup called *Spectroscopy Electron Alpha in Silicon bOx couNter* (SEASON). In a later phase, a transport line to the future DESIR facility [22] is foreseen.

The ion beam detection is performed at multiple locations using micro-channel plate detectors (MCP1,2) at the QMF entrance and exit, and after the PU electrode (MCP3). All 3 MCPs have 10 % transmission grids allowing attenuation of intense beams and also detection of ion currents. The MCP3 detector has an additional phosphor screen for ion beam imaging. Additional detection sites are located around the MR-TOF mass spectrometer [23]. It is also possible to install silicon detectors on linear actuators at the QMF entrance and exit.

The S3-LEB setup with the gas cell, ion guides and PILGRIM mass spectrometer is currently installed in a test room at the LPC Caen institute. All components have been coupled and aligned.

#### *2.2. The GISELE Laser Laboratory*

The purpose of the GISELE laboratory is to perform offline laser ionization and spectroscopy experiments with the elements of interest for the S3-LEB facility. A part of the GISELE laser system has been coupled to the S3-LEB setup at LPC where it is currently being tested. The layout of the full GISELE laser system can be seen in Figure 3. An Nd:YAG laser, working at 10 kHz repetition rate and in the second harmonic, pumps several titanium:sapphire (Ti:sa) lasers with a power distribution achieved by implementing *λ*/2 retardation plates and polarizing beam splitter (PBS) cubes. Intra-cavity and extra-cavity higher harmonic generation can be achieved using nonlinear crystals. The Ti:sa laser beams are overlapped and guided towards an atomic beam unit (ABU) with a high-temperature oven (*Tmax* ∼ 2000 ◦C).

In the future, a dye laser system is foreseen to be implemented to complement the Ti:sa wavelength coverage [24]. Recently, studies of a single-longitudinal-mode, pumped pulsed-dye amplifier have been carried out for high-resolution and high repetition rate spectroscopy applications, when using the dye laser system [25].

**Figure 3.** Typical GISELE laboratory layout. See text for details.

Monitoring and synchronizing the laser pulse time profiles is ensured by picking up a reflection or a fraction of each output laser beam and detecting it with a photodiode, the output of which is connected to an oscilloscope. The temporal overlap of the different

Ti:sa laser beams can be controlled either by modifying the gain using the focusing of the pump light into the crystal, or by Pockels cells. The wavelength is measured using a HighFinesse WS7 wavemeter. A Labview control and acquisition system is used to operate the lasers, record power and wavelength values, count ions in the ABU and perform wavelength scans.

The Z-type Ti:sa cavities of GISELE (see Figure 3) are broadband (BB) cavities, having either a birefringent filter (BRF) plus etalon, or grating as wavelength selective elements and achieving a typical linewidth Δ*flaser*−*f und* of 5–10 GHz of the fundamental output frequency [26,27]. A Z-type cavity using two etalons is available, for achieving a narrower linewidth Δ*flaser*−*f und* of 1.5–2 GHz [28]. For narrowband (NB) spectroscopy, an injectionlocked pulsed-Ti:sa ring laser is available, seeded by an external cavity diode laser (ECDL), achieving linewidths, Δ*flaser*−*f und*, of ≤50 MHz [29]. The ECDL system requires feedback protection, which is provided by optical isolators. Typical output powers with standard 10 W pumping power of these Ti:sa systems are 2.2–2.7 W.

The design of the resonators is optimized so that any astigmatism from the surfaces of the Ti:sa crystal and the curved mirrors at both sides of the crystal cancel each other. The resonator is designed for high repetition rate operation (up to 10–15 kHz).

The ABU consists of an oven, apertures, ion optics and a MCP detector that is kept under vacuum. The atomic beam diffuses in the upward direction and it is collimated by multiple apertures before it reaches the photon–atom interaction region. This helps to minimize the transverse Doppler width of the atomic ensemble, as well as to constrain the interaction volume. To deflect the surface ions, two electrode pairs, located below the photon–atom interaction region, can be biased.

Once ions are created by the photon–atom interaction, an electric field gradient guides the ions towards an MCP located ∼50 cm away from the interaction region. The gradient is optimized in order to obtain a time focus on the MCP allowing mass resolving powers on the order of *R* = 200.

The MCP signal is pre-amplified, then sent to a constant fraction discriminator and, finally, a time to digital converter (TDC) with maximum resolution of 4 ns/bin. The TDC is triggered by a TTL signal synchronized to the Q-switch trigger of the pump laser. The obtained TDC signal is sent to the Labview acquisition system.

#### **3. Results**

#### *3.1. RFQ Offline Tests*

The voltage optimization and the transmission and resolution tests were performed separately for the SRFQ/mRFQ and QMF/buncher. To set the voltages and monitor/control vacuum parameters, a CVI control system with Python interfaces was used. The MCP signals were recorded by a National Instruments 9402 counter and the ion currents by a Keithley 6487 picoampere meter unit.

For the tests of the SRFQ/mRFQ section, a 133Cs source was inserted on a linear actuator in the designed area for the gas jet formation (in front of the SRFQ entrance). The total source current could be measured on a 10% transmission grid covering the source emission area. To achieve the operating pressure in online conditions, argon was injected directly in the gas cell vacuum chamber. The RF driving frequency of the ion guiding RFQs was set to 500 kHz, to allow operation with lower RF amplitudes and avoid discharges. The DC voltages on the mRFQ and SRFQ electrodes were then optimized to enhance transmission. The beam was detected on a Faraday cup placed behind the mRFQ. The transmission tests were performed aiming for the range of background pressures between 10−<sup>2</sup> mbar and 10−<sup>1</sup> mbar, that would correspond to online conditions for the creation of a matched jet of Mach number ∼ 8 by the corresponding nozzles operated at different stagnation pressure regimes.

The optimum SRFQ and mRFQ settings result in a transmission of ≥80(15)% after mRFQ for more than an order of magnitude change in pressure *p*, centered around the region of interest for S3-LEB experiments (see Figure 4). The error bars have been fixed to a 10% value, which is typical beam current uncertainty obtained in our measurements with a picoampere meter. In the same figure a comparison with SIMION simulations [30], using the hard-sphere (HS1) collision model and the same RF, DC and *p* settings, is presented. For these simulations, the ion source was assumed to be a 2*π* emitter from a disk, having 6.5 mm diameter of the used 133Cs source and the energy distribution compatible to the thermal energy of a *T* ≈ 1000 ◦C ensemble. The collisional cross-section *σcol* with argon atoms was estimated from the ionic radius of 133Cs and the Van der Waals radius of argon to be 4.25 × <sup>10</sup>−<sup>19</sup> <sup>m</sup>2. The experiments revealed that the SRFQ and mRFQ have a very high transmission efficiency (75–100(10)%) within the pressure region of interest for creating a matched hypersonic jet of Mach number 7–8. The simulations indicate 60–85% transmission efficiency. The underestimation in the simulations for high pressures can be explained by the limitations of the HS1 collision model or an inaccuracy in the chosen collision cross section. The qualitative trend is nevertheless reproduced well.

**Figure 4.** Experimental and simulated SRFQ and mRFQ transmission efficiency as a function of pressure *p*. Necessary *p* conditions for a matched hypersonic jet of Mach number 7–8 are highlighted by the region of interest in light blue.

The QMF/buncher ensemble was tested with a rubidium surface ion source installed in front of the QMF, providing a mixture of 85,87Rb with the natural abundance. In ion guide mode (no quadrupole DC field), the transmission was close to 100%. When a DC voltage in combination with the RF voltage was applied (filtering mode), the QMF transmission efficiency was checked by a 2D scan of the DC and RF voltages leading to a resolving power on the order of *m*/(2Δ*mFWHM*) ≈ 40 and a transmission of about 40%. For lower resolving powers, the transmission efficiency is above 80%.

In order to give a more explicit estimate of the mass resolving power, a series of scans were performed also while keeping a constant DC to RF voltage ratio, the so-called load– line scan. Knowing the inner radius *r*<sup>0</sup> = 10 mm of the QMF, it was possible to calculate for each RF amplitude the optimal ion mass corresponding to a Mathieu *q* parameter of 0.706 (the tip of the stability diagram). The load–line scan was thus converted into a mass scan, for different DC-RF ratios. In Figure 5, we present one such scan performed with a DC/RF amplitude ratio of 0.166. The mass axis is recalibrated so that the left peak corresponds to 85Rb. This configuration shows a complete separation of 85Rb<sup>+</sup> and 87Rb<sup>+</sup> and allows the possibility of also separating the intermediate mass *A* = 86, with a suppression factor of the side bands, which remains to be determined experimentally. This resolving power is, however, limiting for the separation of heavier masses. With the first production of ions in the gas cell or jet, which will have a different emittance from the beam used in this test, the resolving power figure will be updated. Further improvements can be achieved by a better control of the symmetry of the RF field between the positive and negative phase, which is currently on the order of 1%.

In the same figure SIMION simulations of the QMF transmission are performed with the same settings as in the experiments. The incident ion beam is modeled as a cone matching the diameter of the ion-source collimator of 6 mm and having a half-angle of 2.5◦, which leads to the experimental transmission efficiency through the QMF with a DC voltage of 100 V and optimal RF amplitude (which are the standard settings). One notices that the experimental resolving power is well reproduced.

**Figure 5.** Experimental load–line scan of the QMF for a fixed DC to RF amplitude ratio of 0.166, compared with a SIMION simulation performed with the same parameters. The mass axis is calibrated so that the left peak corresponds to 85Rb.

The transmission through the RFQ*cb* was tested under the same conditions as during the QMF tests, being optimized both in continuous and bunching mode. The helium flow rate injected in the buncher was from 75 to 105 mL/min, the latter being the limit due to the resulting pressure of 1 × <sup>10</sup>−<sup>5</sup> mbar in the PU electrode area, preventing the proper operation of MCP3. However, an increase in flow rate from 75 to 105 mL/min achieved only 25% relative increase in the transport efficiency, making 75 mL/min already close to the optimal pressure. A comparison of ion spots on the phosphor screen showed similar radial distributions between 75 mL/min and 90 mL/min; however, for flow rates < 75 mL/min , a significant degradation of the ion spatial distribution was observed.

The 10% transmission grid on MCP3 was hardwired to the ground potential; therefore, it did not allow us to measure the continuous ion beam through the buncher in continuous mode. For this type of measurement, the beam was collected on the negatively biased PU electrode and read out with the picoampere meter. With the optimum RF, DC and He injection settings, a transport efficiency of the buncher in continuous mode of about 85% was measured on the PU electrode with an uncertainty ∼ 10%. In order to test the buncher in pulsed mode, it was necessary to accelerate the ion bunches to MCP3 using the PU electrode, thus giving them sufficient energy for efficient detection.

The bunched-mode efficiency was tested both in continuous accumulation mode and using a beam gate (BG) to limit the number of ions per bunch and ensure the same cooling time for all ejected ions. A BG was created by switching the injection electrode of the QMF, in order to block the ion beam, with the exception of a short time, controlled by a TTL trigger. The transport efficiency was tested using a BG of 1 ms and a cooling time of 10 ms, leading to a transmission value of 30(10)%. This value was, however, obtained with a low-resolution (50 ns) ion-counting system with an average intensity of one ion per bunch or less. A test with a high-resolution counting system will allow eliminating any potential pile-up effects.

In addition to the transport efficiency, the bunch TOF distribution was recorded using an oscilloscope and its averaging function with 75 mL/min flow rate. This result is presented in Figure 6, left panel, where one observes two overlapping bunches corresponding to the two Rb isotopes already separated in TOF on the MCP3. The fact that the doublepeak structure corresponds to the two isotopes was validated by using the QMF at a DC voltage of 100 V and suitably chosen RF amplitude, to select one or the other isotope. The heights of the two individual peaks were normalized to match the corresponding isotopic abundances.

**Figure 6.** (**Left**) Time-of-flight distribution of rubidium ions behind the RFQ*cb* in three configurations: without any selection from the QMF (green); with QMF selecting 85Rb<sup>+</sup> (blue); with QMF selecting 87Rb<sup>+</sup> (red). The blue and red curves are normalized to 85,87Rb relative abundance. (**Right**) Comparison of the simulated (red) TOF distribution of a 85Rb<sup>+</sup>- 87Rb<sup>+</sup> mixture with 75 mL/min helium flow rate to the experimentally measured one (black).

Simulations of the RFQ*cb* using the SIMION software and the HS1 algorithm were carried out following the same principles as those described for the SRFQ, mRFQ and QMF. Simple conductance calculations knowing the aperture diameters, the pumping power and some of the gauge pressures (corrected for helium) were performed to estimate the true pressure in the buncher. The simulations were started in front of the QMF extraction lens. The ion energy distribution chosen was identical to the one giving the best reproduction of the QMF behavior. For the entire simulation, the experimental voltages were used as input. Two helium flow rates were tested: one set to the experimental value most commonly used (75 mL/min) and one set to a slightly higher value (125 mL/min). The ions were injected all at once, allowed to cool for either 2, 5 or 10 ms, and then extracted towards the MCP3. The simulations showed transmission efficiencies in the experimental pressure range of 20–40%, compatible with the experimental findings.

Furthermore,the simulated TOF distribution of a mixture of 85Rb<sup>+</sup> and 87Rb<sup>+</sup> with the correct elemental abundance was compared to the measurement using the same helium flow rate, and is presented in the right panel of Figure 6. The TOF offset was not measured experimentally with the oscilloscope and thus the simulation TOF was shifted by an arbitrary amount to match the centroid of the experimental spectrum. However, one notes that the experimental width and separation of the peaks is well described.

One must note, however, that all the values described in this section are obtained for the alkali ion source, the emittance (and divergence) of which should be significantly larger than that of the laser-ionized beam.

#### *3.2. Laser Ion Source Offline Tests*

Erbium atoms were chosen for the offline studies based on the fact that during the S<sup>3</sup> commissioning it is planned to use 152Er. The goal of the Er I RIS offline measurements at GISELE is to measure the isotope shift (IS) and hyperfine structure (HFS) by a two step RIS scheme of stable erbium isotopes (164,166,167,168,170Er), and to compare these results with the literature in order to quantify the performance of the equipment and the expected online performance. Stable erbium atoms are deposited in solution form (Er2O3 in 5 % HNO3) on a tantalum foil, which then is placed inside the ABU oven.

The left panel of Figure 7 shows the ionization scheme used in the presented study. The excitation step (415.2 nm) was reported in [31] and, recently, precise Rydberg and auto-ionizing state, and ionization potential measurements, were carried out, starting from the same level at 24,083.2 cm−<sup>1</sup> [32]. From the latter work, the most efficient A.I. state transition of 25,210.4 cm−<sup>1</sup> was chosen for the ionization energy. Moreover, the strength of the excitation step has been determined to be A*ki* = 9.6 × <sup>10</sup><sup>7</sup> <sup>s</sup>−<sup>1</sup> [33].

**Figure 7.** (**Left**) Er I two-step ionization scheme used for NB RIS measurements [32]. On the left hand side of the diagram, the excited state, the ionization potential (I.P.) energy and the populated auto-ionizing (A.I.) state levels are presented, on the right hand side, electron configuration and total angular momentum *J* are shown. (**Right**) TOF spectrum of the Er ions observed with the NB Ti:sa system using the scheme shown in the left panel.

In these measurements, the NB Ti:sa system with a fundamental output linewidth 20 ≤ Δ*f* ≤ 50 MHz was used for the excitation step and a BB Z-type Ti:sa cavity was used for the ionization step. The ABU TOF resolution with stable erbium atoms was *R* = TOF/(2 × FWHM) ∼ 260, with TOF = 21.4 μs and FWHM170Er = 40 ns. An acquired TOF spectrum resolving all stable Er isotopes following their natural abundances is shown in the right panel of Figure 7.

The wavelength adjustment of the excitation step was performed using the Labview control and acquisition system, which adjusts the ECDL master laser output wavelength. For each scan step the corresponding TOF spectra is saved. An individual resonance of each isotope can then be extracted from the full TOF spectra by choosing a region of interest.

After the frequency doubling stage, once the NB Ti:sa system beam reached the ABU, the measured full power after the two ABU windows varied between 30 and 100 mW. The Z-type Ti:sa BRF cavity used intra-cavity second harmonic generation and produced 40–100 mW of power at the ABU.

Before the IS and HFS measurements, the saturation power level *P*<sup>0</sup> of the excitation step was measured. In these measurements, both lasers were on resonance and the ionization step was kept at full power. Neutral density filters were used to reduce the laser power. The spatial alignment of both beams was performed by using the TDC count rate and a pair of ABU entrance/exit window apertures. The results are represented in Figure 8. The data set has been fitted by using the following equation:

$$I(P) = A + \mathbb{C} \times (P/P\_0)/(1 + (P/P\_0))\_\prime \tag{1}$$

with *A*, *C*, *P* and *P*<sup>0</sup> being an offset describing influence from surface- and non-resonant ionization, the maximum resonant ionization rate, measured power and saturation power, respectively. The fit results were: *A* = 0.5(10) cps, *C* = 110(10) cps. The extracted saturation power *P*<sup>0</sup> was 145(40) μW. The beam spot diameter was about 1 mm.

Moreover, to observe the saturation effect more precisely, scans at several excitation step powers *P* were performed. In the measurements shown here the power was reduced until no more influence on the full-width at half-maximum (FWHM) was observed. This was the case at about 10–20 μW resulting in a resonance linewidth Δ*fres* of ∼120 MHz. The

expected natural linewidth is ∼15 MHz. The saturation power from previous work [32] was 2.1(1) mW, compared with the present result of 0.145(40) mW. The reduction of the saturation power in our case can be explained by the reduced linewidth of the NB Ti:sa system (20 ≤ Δ*f* ≤ 50 MHz) in comparison to the BB Z-type Ti:sa cavity (Δ*f* ∼ 5 GHz) and by possible differences in beam spot diameter used in [32].

IS and HFS measurements were performed for the different stable erbium isotopes. A detailed analysis of the data will be presented in a forthcoming paper [34], where the IS and HFS parameters will be represented.

Fitting of the raw data was carried out by a *χ*<sup>2</sup> procedure in SATLAS [35]. An IS result from a single scan of 166,170Er is presented in the left panel of Figure 9. The RIS measurements were performed with 10–20 μW power levels for the excitation step and at 20–90 mW for the ionization step.

**Figure 8.** Measured count rate *I* as a function of excitation step power *P* for the RIS scheme presented in Figure 7. The orange curve represents a fit to the measured data.

**Figure 9.** (**Left**) Normalized IS measurements of <sup>166</sup>−170Er I (red/orange curve—SATLAS [35] *χ*<sup>2</sup> fit of the data; *f*<sup>0</sup> = 721.9966 THz; excitation and ionization step powers are represented in the text box). (**Right**) Scattering of individual IS (Δ*f*) measurements around the weighted average IS (Δ*fWA*) from all NB RIS measurements. The used RIS scheme is presented in Figure 7.

A scatter of the IS data from 20 measurements is presented in the right panel of Figure 9, with the weighted average subtracted from all values. The individual uncertainties of the data points represent statistical uncertainties, multiplied by *χ*2 *red* to correct for non-statistical scattering effects. The source of the larger data scattering is still under investigation.

Owing to the narrow spectral linewidth of the NB Ti:sa system, the HFS spectra of the odd–even 167Er isotope could also be measured. The total angular momentum of the ground state (g.s.) *Jg*.*s*. = 6 and nuclear spin *I* = 7/2, results in 8 g.s. HFS components ranging from *F* = 5/2 to 19/2. The excited state (e.s.) has angular momentum of *Je*.*s*. = 5, also with 8 HFS components ranging from *F* = 3/2 to 17/2 (all *J* values taken from [33]). By applying selection rules, this results in 21 possible transitions. The splitting of the g.s. components has been measured by A. Frisch et al. [36]. The e.s. hyperfine constants *A* and *B* are unknown.

The HFS information was extracted from 11 scans, performed below the saturation power level. The g.s. *Al* and *Bl* coefficients were fixed to the the literature values −120.487(1) and −4552.984(10) MHz [36], respectively. The nuclear magnetic octupole moment coefficients *Cl* and *Cl* for the g.s. and e.s. were set to 0. The fit result for a single scan is represented in Figure 10. The spectrum corresponding to all 21 HFS components has been recorded and fitted. In the presented fitting procedure, the peak intensities are left as free variables.

**Figure 10.** Normalized single HFS measurement of 167Er I with an inset in the top left corner containing details of the least intense/clearly resolved HFS components *a* − *f* (red/green curve— SATLAS [35] *χ*<sup>2</sup> data fit/atomic resonance positions based on input parameters; the weakest peaks according to atomic theory have been magnified for visualization purpose and are presented in the insets, with a multiplication factor added to the HFS component; *f*<sup>0</sup> = 721.9966 THz; text box presents the used excitation and ionization step powers *Pex* and *Pion*; the used RIS scheme is presented in Figure 7).

#### **4. Outlook and Conclusions**

The commissioning of the S3-LEB setup is entering the offline test phase of the entire installation, in which the the gas cell, RFQ chain and the MR-TOF mass spectrometer are connected and the laser system is coupled to the gas cell.

The commissioning tests performed separately for the RFQ tandems of the setup (SRFQ/mRFQ and QMF/RFQ*cb*) have shown promising results, both in terms of transmission and resolving power/bunching capability. Work is ongoing with the cooling and bunching section to improve the performance before the first ion injection into the MR-TOF mass spectrometer will take place.

The Ti:sa-based GISELE offline laser laboratory at GANIL has been successfully developed for the high-resolution spectroscopy requirements of S3-LEB. The laser systems are adapted for both in-gas-cell and in-gas-jet spectroscopy methods. Using one of the possible Er I RIS schemes, new narrowband IS measurements of 164,166,168−170Er have been performed, and the stability of the system between different measurements has been illustrated. With the same RIS scheme, first high-resolution HFS spectra with stable 167Er has been measured.

**Author Contributions:** Conceptualization, J.R., A.A., M.A., F.B., L.C., J.-F.C., A.C., S.D., P.D. (Pierre Delahaye), P.D. (Philippe Desrues), A.D., P.D. (Patricia Duchesne), R.F., X.F., S.F., P.G., R.P.d.G., S.K., N.L., R.L., J.L., F.L., V.M., Y.M., I.M., A.O.-C. , B.O., J.P., O.P., B.-M.R., H.S., S.S., E.T., J.U., C.V., M.V. (Marine Vandebrouck), P.V.d.B., M.V. (Matthias Verlinde), E.V. and K.W; methodology, J.R., A.A., P.D. (Pierre Delahaye), R.F., X.F., S.F., V.M., P.V.D. and A.O.-C.; software, V.M. and C.V.; validation, J.R., A.A., L.C., R.F., X.F., S.F., N.L., V.M., H.S., S.S., P.V.D. and A.O.-C.; formal analysis, J.R., A.A., V.M. and A.O.-C.; investigation, J.R., L.C., R.F., X.F., N.L. and V.M.; resources, R.F., N.L. and P.V.D.; data curation, J.R., A.A., V.M. and A.O.-C.; writing—original draft preparation, J.R.; writing review and editing, L.C., R.P.d.G., R.F., X.F., N.L., V.M., I.M., H.S., P.V.D. and K.W.; visualization, J.R. and V.M.; supervision, N.L., R.F. and P.V.D.; project administration, N.L., R.F., V.M. and P.V.D.; funding acquisition, N.L. and P.V.D. All authors have read and agreed to the published version of the manuscript.

**Funding:** S<sup>3</sup> has been funded by the French Research Ministry, National Research Agency (ANR), through the EQUIPEX (EQUIPment of EXcellence) reference ANR-10EQPX- 46, the FEDER (Fonds Europeén de Developpement Economique et Regional), the CPER (Contrat Plan Etat Re ´ gion), and ´ supported by the U.S. Department of Energy, Office of Nuclear Physics, under contract No. DE-AC02-06CH11357 and by the E.C.FP7-INFRASTRUCTURES 2007, SPIRAL2 Preparatory Phase, Grant agreement No.: 212692. S3-LEB: This project has received funding from the French Research Ministry through the ANR-13-B505-0013, the Research Foundation—Flanders (FWO)—under the International Research Infrastructure program (nr. I002219N), the Research Coordination Office—KU Leuven—-the European Research Council (ERC-2011-AdG-291561-HELIOS) and the European Union's Horizon 2020 research and innovation program under grant agreement No 654002.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** This setup results from the collaborative work of the IGLIS newtork, grouping many research centers and universities such as CEA-Saclay (IRFU), CERN (CRIS), GANIL, IBS-RISP, IJCLab, IMP, JAEA, Johannes Gutenberg-Universität Mainz (Institut für Physik/LARISSA), JINR (GALS), JYFL (IGISOL/MARA), KEK (KISS), KU Leuven, MSU, Nagoya University, Normandie Université (LPC Caen), Peking University, RIKEN (SLOWRI/PALIS), TRIUMF (TRILIS), Université de Strasbourg (IPHC), University of Manchester, University of Tsukuba and Laboratoire de Physique des 2 infinis Irène Joliot-Curie (IJCLab) (for more details about IGLIS collaboration please refer to our network page [37]).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

