Laser spectroscopy with an electrostatic ConeTrap

This is an electronic reprint of the original article. This reprint may differ from the original in pagination and typographic detail. Author(s): Title: Year: Version: Please cite the original version: All material supplied via JYX is protected by copyright and other intellectual property rights, and duplication or sale of all or part of any of the repository collections is not permitted, except that material may be duplicated by you for your research use or educational purposes in electronic or print form. You must obtain permission for any other use. Electronic or print copies may not be offered, whether for sale or otherwise to anyone who is not an authorised user. Laser spectroscopy with an electrostatic ConeTrap Kelly, S.; Campbell, P.; Cheal, B.; Eronen, Tommi; Geldhof, Sarina; Jokinen, Ari; Moore, Iain; Penttilä, Heikki; Pohjalainen, Ilkka; Rinta-Antila, Sami; Sonnenschein, Volker; Voss, Annika Kelly, S., Campbell, P., Cheal, B., Eronen, T., Geldhof, S., Jokinen, A., Moore, I., Penttilä, H., Pohjalainen, I., Rinta-Antila, S., Sonnenschein, V., & Voss, A. (2017). Laser spectroscopy with an electrostatic ConeTrap. Hyperfine Interactions, 238(1), Article 42. https://doi.org/10.1007/s10751-017-1412-z 2017 Hyperfine Interact (2017) 238:42 DOI 10.1007/s10751-017-1412-z Laser spectroscopy with an electrostatic ConeTrap S. Kelly1·P. Campbell1·B. Cheal2·T. Eronen3·S. Geldhof3·A. Jokinen3· I. D. Moore3·H. Penttil¨ a3·I. Pohjalainen3·S. Rinta-Antila3· V. Sonnenschein3·A. Voss3 © The Author(s) 2017. This article is published with open access at Springerlink.com Abstract A compact electrostatic trap has been designed and installed as part of the recent upgrades to the IGISOL IV facility. The ConeTrap provides an in vacuo optical pumping site for low energy (800 eV) ionic ensembles available for interaction periods of 10-100 ms. At present, 6.7(3) % of injected mass A=98 ions can be trapped, stored for 5 ms, extracted and transported to a laser-ion interaction region. This fraction represents those ions for which no perturbation to total energy or energy spread is observed. Proposed enhancements to the trap are designed to improve the trapping efficiency by up to a factor of 5. Differential pumping and reduction in background pressure below the present 10−6mbar will extend storage times beyond 100 ms. Keywords Electrostatic ·Ion ·Trap ·Laser ·Spectroscopy This article is part of the Topical Collection on Proceedings of the 10th International Workshop on Application of Lasers and Storage Devices in Atomic Nuclei Research: “Recent Achievements and Future Prospects” (LASER 2016), Pozna´ n, Poland, 16–19 May 2016 Edited by Krassimira Marinova, Magdalena Kowalska and Zdzislaw Błaszczak S. Kelly [email protected].ac.uk B. Cheal [email protected] 1Nuclear Physics Group, Schuster Laboratory, Brunswick Street, University of Manchester, Manchester M13 9PL, UK 2Oliver Lodge Laboratory, University of Liverpool, Liverpool L69 7ZE, UK 3JYFL, University of Jyv¨ askyl¨ a, Jyv¨ askyl¨ a, Finland 42 Page 2 of 11 Hyperfine Interact (2017) 238:42 1 Introduction At the laser-IGISOL facility JYFL, there has been an increased, experiment driven, desire for a field free, high vacuum, trapped ion interaction region in which to perform long interaction (10-100 ms) laser spectroscopy. A ConeTrap, pioneered by Schmidt et al. [1], would provide such a location. These electrostatic traps can successfully contain 30 000 ions for in excess of 100 ms [1,2], and are thus ideal for the purpose of laser spectroscopy. ConeTraps comprise of two (or more) conical, reflecting electrodes that are used to provide multi-reflection based containment of ions. Simple conical electrodes at voltage create potentials which internally are harmonic in both the zand rdirections (specified in cylindrical coordinates). When two facing cones are partitioned with a grounded aperture electrode then electrostatic potentials such as those shown in Fig. 1can be formed. It is observed that, even with extended central sections, stable ion trapping can be achieved (Fig. 2) for a range of potentials and various reflections in the ConeTrap. The ions in the ConeTrap have a maximum axial velocity in the central region, determined by the electrode voltages. Each ion will momentarily have a net axial velocity of zero at each end of the trap electrodes. These extremes, the turning points and the trap centre, are associated with extremes in Doppler-shifted frequency where optical resonance can be induced in trapped ions. In the centre, velocity spread compressed, collinear laser spectroscopy can be attempted, albeit at modest total energy (800 eV). At the turning points a laser-ion interaction perpendicular to the direction of motion is achieved irrespective of energy (given a sufficiently narrow laser beam). 2 Description and basic operation of the ConeTrap Initially, the ConeTrap implemented at the IGISOL IV laboratory was comprised of two conical electrodes separated by a grounded central chamber, following the designs outlined in reference [3]. The conical electrodes were made with open ends, allowing laser light to pass through the ConeTrap and interact with ions in the central region. The restricted stability and critical alignment dependencies led to design changes that resulted in the removal of the grounded central region and development of a ‘collapsed’, two cone only, trap. The final design is highly compact and electrostatic simulations (described below) reveal broad stability regions achievable in a range of trapping modes. The device is shown in Fig. 3. The ConeTrap is constructed from high purity (deoxygenated) copper and PEEK insulators, with the stainless steel supporting rods and meshes external to the trap. The copper electrodes are each 74.0 mm in length, the smaller aperture, φ1= 23.0 mm and the larger aperture, φ2=34.0 mm, creating an opening angle of 8.46◦. The two electrodes are separated by 4.0 mm of insulation and a grounded central aperture 2.0 mm thick with opening diameter 8.0 mm. At the IGISOL, the ConeTrap is situated on the high voltage platform that houses the RFQ cooler. At this point, ions have a kinetic energy of approximately 800 eV, thus moving with a lower velocity than any other section of the high vacuum beamline. Only modest trapping potentials and fast switching (50 ns) of voltages less than ∼2 kV are required to operate the device. When positioned on one face of a switchable quadrupole bend, a laser-ion interaction region covering the full injection path of the ConeTrap can readily be achieved. To enable ions to enter the trap the front electrode is switched to an ‘injection’ voltage and then back to a ‘trapping’ voltage, prior to the reflected bunch returning to the front cone. Hyperfine Interact (2017) 238:42 Page 3 of 11 42 800 V: 675V: (d) (a) (b) (c) Fig. 1 Plots showing equipotential contours for: (a) Front cone at injection/extraction potential (680 V) and rear at 805 V, (b) symmetric ConeTrap with both electrodes at 805 V, (c) asymmetric ConeTrap with the front and rear electrodes set to 1350 V and 805 V respectively. Also included is, (d) a 3-D projection of the harmonic potential created in the zdirection for the asymmetric regime, as shown in (c) (a) Soft reflection (b)Hard reflection Fig. 2 Trajectories of ions within the ConeTrap in two different trapping modes After a desired trapping time, switching down the front electrode permits extraction and secondary acceleration. At the IGISOL, the ions that exit the trap are guided to the ‘Light Collection Region’ (LCR) in the collinear laser beamline. In the absence of collisions with residual gas, or ion-ion interactions, an ionic ensemble can be indefinitely trapped in either of two symmetric trapping modes, denoted soft and hard reflections here. A soft reflection is established, at Eion= 800 eV, when both electrodes are set to a voltage of ∼810 V. Hard reflection is achieved at potentials of 1250–1400 V. It is also possible to achieve stable modes using an asymmetric combination of these reflections. The soft reflection results in ions spending significant time, >10 μs at A=100 and 800 eV, transiting within the electrode compared to that, <2μs, characteristic of a hard reflection. A 42 Page 4 of 11 Hyperfine Interact (2017) 238:42 (a) (b) Fig. 3 Two depictions of the ConeTrap installed at the IGISOL IV beamline. Panel a) shows the electrode structure as in SIMION [4], panel b) shows the manufactured component, plus fore-line ion injection Faraday cage soft reflection in the back cone and a hard reflection in the front causes ions to spend a significantly longer time away from the switching potential of the front electrode. An asymmetric ConeTrap, with respect to potentials, provides the optimal solution for injection and extraction through the same electrode, and was employed in all testing of the ConeTrap. 3 Intra trap optical pumping For 800 eV ensembles with energy spreads of 0.6 eV, typical for the JYFL coolerbuncher [5], reduced Doppler widths of <100 MHz would be expected for A=100 ions at 300 nm. An efficient ConeTrap thus provides an ideal interaction region for optical pumping of transitions in ionic species. At the IGISOL such optical pumping has already been used to permit spectroscopy of, previously inaccessible, short-lived radio isotopes [6,7]. In references [6,7] pumping was performed within the gas-filled cooler buncher and was limited in linewidth by the macro motion within the RFQ trap (corresponding to 10’s of GHz of line broadening). More critically, the spectroscopy was limited in the excitation energy of the populated excited state due to non-photonic relaxation processes in the trapped volume. A ConeTrap would readily overcome these limitations and permit post trapping spectroscopy if extracted ensembles could be recovered with minimal perturbation of the energy and energy spread. Each has been explored at the IGISOL and the results are discussed here. 4 ConeTrap parameters Full simulation of the ConeTrap has been conducted using the ion-optic simulation package SIMION [4], in which the potential is initially evaluated (in cylindrical geometry) by numerically solving the Laplace equation for finite grid elements (with 1 mm resolution). Particle trajectories are then evaluated using a fourth order Runge-Kutta method in three dimensions [8]. Particle-particle and particle-gas interactions were neglected in the simulations (but explored experimentally). Figure 4shows the electrostatic model of the post-cooler environment (referred to as a workbench in SIMION). The behaviour of ions can be modelled, and the survival as well as energy and temporal spread predicted for the full trapping period. The simulation was set such that the ConeTrap would trap the ions after a user defined time, replicating realistic operating conditions including temporal and energy spread of the Hyperfine Interact (2017) 238:42 Page 5 of 11 42 Fig. 4 The SIMION workbench for the ConeTrap and all post-cooler elements used in guiding and focussing the ion beam. Shown in the right hand corner is a cross section of the asymmetric ConeTrap employed at typical operating conditions Fig. 5 Stability plot for the collapsed ConeTrap, showing the survival percentage of ions that are contained for 5 ms, then released to the LCR (as modelled in SIMION). Negligible losses to buffer gas or ion-ion collisions are assumed bunched ion plume. ConeTrap parameters that maximise the trapping efficiency were found, as functions of the trapping, rear and injection voltages. Injection and extraction voltages were found to optimise at the same value, reducing the complexity encountered in switching potentials in the ConeTrap. The region shown in Fig. 5focuses on rear cone voltages of 800 – 810V, corresponding to >6μs transit in the rear cone (the minimum storage time for the efficient capture of ion plumes from the IGISOL cooler-buncher) and extend to 1400 V for the front cone beyond which trap stability is rapidly lost. The observation of a relatively 42 Page 6 of 11 Hyperfine Interact (2017) 238:42 narrow stability region near symmetric potentials and broader stability region for higher potentials (separated by a region of instability) closely follows the behaviour predicted and observed in other ConeTraps [3]. 5 On-line results Experimental testing of the collapsed ConeTrap guided by the SIMION modelling was undertaken at a range of masses and trapping potentials. With the trapping time of the ConeTrap set to 5 ms, the released plumes were deflected to micro channel plates (MCPs) at the LCR (and viewed on an oscilloscope). To ensure count rates free of saturation, ion fluxes were limited to 0 – 5 ions per bunch. Ion plumes arriving 5 ms after cooler-buncher release were evident as (shown in Fig. 6). The temporal spread of the ion plume no longer reflects the 10 –15 μs characteristic release of the JYFL cooler buncher [9] and instead reflects the cycle time in the ConeTrap, dominated by the orbit time in the rear cone electrode. Figure 7shows that the bunch width decreases with an increasing rear voltage in a manner entirely consistent with the SIMION simulation predictions. At rear voltages of 805 – 808 V >60 % of the typical ion plumes can be captured by the device (Vfront=1350 V). Measurements of trapping efficiencies were made using radioactive ions around mass A=100. The ions were produced in fission using 25 MeV protons incident upon a natural uranium target. At mass A=98 a range of strontium, yttrium and niobium systems could be investigated by directly monitoring decay gamma emission with and without storage in the ConeTrap (overcoming saturation issues with the channel plates). A direct comparison between count rates with and without storage in the ConeTrap were made using spectra such as those shown in Fig. 8. The ConeTrap extraction efficiency showed no dependence on chemical species and could be estimated using data from multiple photopeaks (highlighted in Fig. 8) to be 6.7(3) % survival from cooler-ejection to end of laser line detection. Whilst clearly demonstrating that the ConeTrap is operational the result, an order of magnitude less than our expected (simulated) maximum, suggests significant losses and potentially significant improvements can be made to the device (Section 6). Collinear laser spectroscopy was performed on re-accelerated ionic ensembles (previously stored in the ConeTrap) under off-line conditions. A full a description of the laser spectroscopy experimental procedure can be found in reference [10]. For these tests the 180Hf ion was studied (which can be efficiently explored on transitions from its ionic ground state without optical pumping). The 0 – 33180.92 cm−1resonance in 180Hf was inspected using direct bunched beam spectroscopy and compared to that observed following 5 ms storage in the ConeTrap. The spectra are displayed in Fig. 9. A total of 197 scan (corresponding to 1.955 s/channel) were taken with ions bypassing the ConeTrap, while 2007 scans (corresponding to 30.1 s/channel) were recorded for ion ensembles having been stored for 5 ms in the device. The linewidth of the resonances (Fig. 9) corresponding to those of trapped ions and those guided directly to the LCR were found to be 29(6) MHz and 35(3) MHz respectively. No shift in centroid within errors was detected, suggesting that the mean energy of the transported bunch had remained constant. The possibility that only an energy analysed fraction of the total ensemble has been transported to the LCR can only be excluded once higher overall injection-storage-extraction efficiency has been achieved (and the linewidth and centroid investigated at those conditions). At present only the 6.7(3) % transported fraction can be confirmed to maintain energy and energy spread. Hyperfine Interact (2017) 238:42 Page 7 of 11 42 Fig. 6 Ion plumes arriving at 5 ms after cooler release signal Fig. 7 Plot comparing the measured and simulated temporal plume width on release and the transit time in the rear electrode 6 Improving the yield To improve on the 6.7(3) % efficiency presently achieved on-line further, critical, focussing ion optics are needed. Simulations run at the experimentally determined optimum trap parameters (with realistic plume widths of 10 μs and 0.6 eV energy spread) suggest the 42 Page 8 of 11 Hyperfine Interact (2017) 238:42 0 500 1000 1500 2000 0 1000 2000 3000 4000 5000 Energy (keV) Counts Nb:787 keV Sr: 119 keV Y: 1223 keV Y(Isomer): 1591 keV (a) (b) 0 500 1000 1500 2000 0 100 200 300 400 Energy (keV) Counts strontium: 119 keV Y: 1223 keV Nb:787 keV Y(Isomer): 1591 keV Fig. 8 Germanium detector spectra from A=98 fission fragment isobars detected at the LCR collected over 300 s (peaks highlighted in red indicate those used in storage efficiency estimates) with the (a) ConeTrap bypassed and (b) 5 ms ConeTrap storage ConeTrap to presently optimise close to the edge of trap stability and at parameters that do not provide efficient post-trap transport conditions. Significant losses are predicted to occur at the first quadrupole switching bend (see Fig. 10) and changes to ion optical elements have been explored. The full workbench simulation suggests that an additional applied Einzel lens at the injection/extraction electrode can simultaneously permit both optimised trapping efficiency and optimised downstream transport. Figure 10 shows the suggested optical changes and ray traces at optimised conditions. Following commissioning of the new ion-optics further optimisations including faster switching HV supplies will be explored. During these tests it is intended that single in-out