Development of a resonant laser ionization gas cell for high-energy, short-lived nuclei
T. Sonoda, M. Wada, H. Tomita, C. Sakamoto, T. Takatsuka, T. Furukawa, H. Iimura, Y. Ito, T. Kubo, Y. Matsuo, H. Mita, S. Naimi, S. Nakamura, T. Noto, P. Schury, T. Shinozuka, T.Wakui, H. Miyatake, S. Jeong, H. Ishiyama, Y. X. Watanabe, Y. Hirayama, K. Okada, A. Takamine
aa r X i v : . [ phy s i c s . i n s - d e t ] O c t Development of a resonant laser ionization gas cell for high-energy, short-livednuclei
T. Sonoda a , ∗ , M. Wada a , H. Tomita b , a , C. Sakamoto b , a , T. Takatsuka b , a , T. Furukawa c , a ,H. Iimura d , a , Y. Ito c , a , T. Kubo a , Y. Matsuo a , H. Mita e , a , S. Naimi a , S. Nakamura e , a , T. Noto b , a ,P. Schury e , a , T. Shinozuka f , T. Wakui f , H. Miyatake g , S. Jeong g , H. Ishiyama g , Y.X. Watanabe g ,Y. Hirayama g , K. Okada h , a , A. Takamine i , a , a RIKEN Nishina Center for Accelerator-Based Science, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan b Faculty of Engineering, Nagoya University, Nagoya 464-8603, Japan c Department of Physics,Tokyo Metropolitan University, Tokyo 116-8551, Japan d Japan Atomic Energy Agency (JAEA), Tokaimura 319-1100, Japan e Department of Physics, Tsukuba University, Tsukuba 305-8577, Japan f Cyclotron and Radioisotope Center, Tohoku University, Sendai 980-8578, Japan g High Energy Accelerator Research Organization (KEK) 305-0801, Japan h Department of Physics, Sophia University, 7-1 Kioicho, Chiyoda, Tokyo 102-8554, Japan i Department of Physics, Aoyama Gakuin University, 4-4-25 Shibuya, Shibuya-ku, Tokyo 150-8366, Japan
Abstract
A new laser ion source configuration based on resonant photoionization in a gas cell has been developed at RIBF RIKEN. Thissystem is intended for the future PArasitic RI-beam production by Laser Ion-Source (PALIS) project which will be installed atRIKEN’s fragment separator, BigRIPS. A novel implementation of differential pumping, in combination with a sextupole ion beamguide (SPIG), has been developed. A few small scroll pumps create a pressure difference from 1000 hPa ∼ − Pa within ageometry drastically miniaturized compared to conventional systems. This system can utilize a large exit hole for fast evacuationtimes, minimizing the decay loss for short-lived nuclei during extraction from a buffer gas cell, while sufficient gas cell pressure ismaintained for stopping high energy RI-beams. In spite of the motion in a dense pressure gradient, the photo-ionized ions insidethe gas cell are ejected with an assisting force gas jet and successfully transported to a high-vacuum region via SPIG followed bya quadrupole mass separator. Observed behaviors agree with the results of gas flow and Monte Carlo simulations.
Key words:
Laser ion source, Gas jet, Resonant laser ionization, Laser spectroscopy
PACS:
1. Introduction
Radioactive ion beam (RIB) facilities based on the in-flight production technique provide a wide variety of exoticnuclei without restrictions on lifetimes or chemical proper-ties. An essential requirement for present and future RIBfacilities is to transform this high-energy beam into a low-energy, low-emmitance beam. Such low-energy beams openup opportunities to study ground state properties of ex-otic nuclei by experimental techniques such as laser spec-troscopy and ion trapping. At RIKEN, a universal slow RI- ∗ Corresponding author. Tel.: +81-48-467-8640; fax: +81-48-462-7563.
Email address: [email protected] (T. Sonoda). beam facility, SLOWRI, based on a gas catcher cell withan RF-carpet ion guide [1][2], was assigned as one of theprincipal facilities of RIBF. A novel method, named PALIS(PArasitic RI-beam production by Laser Ion-Source) [3]was also approved for the construction, to expand the us-ability and reduce experimental costs by utilizing unusedRI-beams produced by projectile fragmentation or in-flightfission. Using this scheme, it will be possible to performlow-energy RI-beam experiments alongside every on-lineBigRIPS experiment.At RIKEN RIBF, the RI-beams of highly exotic nucleiare available with the highest intensity in the world. How-ever, the usability is restricted to short periods of beamtime due to high demand and limited yearly operating
Preprint submitted to Elsevier 27 July 2018 ours due to the electrical cost for accelerator operation.To bring about the most effective utilization of such high-performance facility, parasitic production of unused RI-beams would be valuable. In-flight fission and fragmenta-tion produces a beam which is a mixture of thousands ofisotopes. A fragment separator selects one specific RI-beamby removing the vast majority of these isotopes. The re-moved isotopes still include many rare nuclei of interest fornuclear studies; wastefully, they are simply thrown away.Our method is to save these rare isotopes before their re-moval in a beam purification slit. By installing a gas catchercell in the vicinity of the slit at the F1 or F2 focal plane inBigRIPS [4], the RI which would have been removed by theslit can be collected and salvaged as a low-energy RI-beam.A schematic view of this setup is shown in Fig. 1. Due tospace and accessibility constraints, a big gas cell such as theones typically used in RIB facilities [1][5][6] is not possible.Instead, we must use a compact cell with a simpler mech-anism. Such a compact gas catcher cell necessitates higherpressure in order to maintain enough stopping power forthe high energy RI-beams. Therefore, we will use a laserionization gas cell [7]-[10] which can be a few hundred cm in volume with 1 bar argon gas. The thermalized ions arequickly neutralized in high-pressure argon gas and trans-ported by gas flow toward the exit of the cell, where theyare selectively re-ionized by resonant laser radiations in thevicinity of the exit. They can then be further purified byan electro-magnetic mass separator and transported to thelow-energy experimental room. In this way, the parasiticlow-energy RI-beams could be delivered whenever BigRIPSexperiments are in operation. Fig. 1. Schematic of the laser ionization gas catcher setup. Withoutdisturbing the main beam, an unwanted portion of the beam isstopped in the argon gas cell. Neutralized atoms are transported bygas flow towards the exit and re-ionized by resonant laser radiations.Thusly produced low-energy RI-beams can be utilized for slow-RIbeam experiments.
In order to provide a high stopping efficiency for high-energy RI-beams, the pressure in the gas cell must be suffi-cient to ensure the stopping range is shorter than the finitelength of the gas cell. Additionally, the transport of neutral atoms by gas flow constrains the volume of the gas cell dueto diffusion loss. Thus, it is optimal to use the highest pos-sible pressure to allow the smallest possible gas cell. Addi-tionally, a fast evacuation time is necessary to avoid decaylosses for short-lived nuclei during transportation inside thegas cell. This can be accomplished by enlarging the size ofthe exit hole. For dealing with such high gas throughput,while keeping a high-vacuum extraction beam line, largeroots pumps with pumping speed of the order of 10 m /hare typically used in a differential pumping system. The ra-diofrequency sextupole ion beam guide (SPIG) [11]-[14] canmoderate the pumping load by decreasing the conductancebetween the gas cell and the extraction chamber. Even us-ing such high-throughput pumps and SPIG, however, thereare limits to the allowable pressure and consequently thenumber of feasible RIs extracted from the gas cell due to alack of stopping efficiency and a relatively slow evacuationtime. In order to address such inherent limitations, whilealso miniaturizing the entire system to fit within the highlyconstrained space limitations of the PALIS installation, wehave developed a new idea which is the stepwise differen-tial pumping method by small pumping capacities. Thisenables use of an even larger exit hole, more than 1 mm ina diameter, with the use of pressures up to one atmosphereargon in the gas cell. For such a pressure and exit hole, RI-beams with energies up to 10 MeV per nucleon will havea stopping range within the length of the gas cell (25 cm),while the evacuation time of the cell (gas cell volume ∼ ) will allow a half-life of 100 ms to be extracted with20% efficiency.To provide proof of principle for this new gas cell basedlaser ionization system, a prototype gas cell and a beam ex-traction system has been built for off-line experiments. Thedifferential pumping capability of this new system has beenverified; a pressure difference from 1000 hPa argon in thegas cell down to 10 − Pa at the quadrupole mass filter hasbeen achieved, while using a 1 mm diameter gas cell exithole. This is the first result obtained in this mode: resonantlaser ionization inside the gas cell, along with ion extrac-tion from the gas cell and transport to high-vacuum. Thishas been experimentally confirmed off-line for several sta-ble elements produced by filament evaporation or ablationby YAG laser. Resonant ionization has been performed us-ing a two-step excitation scheme with an excimer-dye-lasercombination. The extraction time profiles of the gas celland the SPIG have been investigated and also examinedby a Monte Carlo simulation combined with a gas flow cal-culation. The experimental results show that ions can betransported from a high pressure to a high-vacuum regionvia a long SPIG (253 mm) with an assisting force gas jet.This technique will relieve the inherent restrictions of con-ventional Ion Guide Isotope Separator On-Line (IGISOL)technique [15] based gas catcher cells, providing faster ex-traction times and improved stopping efficiency.2 . Prototype gas cell and a beam extraction system
Fig. 2. Cut-away view of the prototype gas cell and the differentialpumping system for PArasitic RI-beam production by Laser Ion–Source (PALIS).
A prototype gas cell and a beam extraction system hasbeen developed. A conceptual sketch is shown in Fig. 2. Thesystem is composed of four parts: a laser ionization gas cell,a differential pumping system, a quadrupole mass separa-tor and a detector station comprising a Channel Electron-Multiplier (CEM). In this system, differential pumping –from the high pressure gas cell to the ion detector in ahigh-vacuum region – is achieved in just 1 m by applyingour novel differential pumping method without using largeroots pumps.2.1.
The gas cell
In the present off-line setup, a gas cell consisting of asimple cross chamber with dimension of 120 × φ
70 mmwas used. It has an exit hole of 1 mm diameter and twofeedthroughs to provide an electric current to a filament.In order to create a laminar gas flow inside the cell, theshape of the gas inlet was carefully fabricated by referencingflow simulation results. The gas inlet structure is shown inFig. 3. The buffer gas is introduced via an inlet pipe of 6.35mm diameter and spread with the conical structure. Laserbeams can be introduced into the ionization region insidethe gas cell via a quartz view port. Two alternative laserpaths, transverse and longitudinal to the beam extractionaxis, exist for ionization inside the gas cell. In the presentexperiment, however, only the transverse path was used.In the off-line setup, there are two methods available forproducing atomic vapor inside the gas cell: evaporation ofa filament or ablation by YAG laser. For laser ablation,the filament was replaced by a target which was irradiatedby a YAG laser via a view port in the flange opposite thetarget. The gas handling system was specially designed toremove any impurity or contaminant from the argon buffer gas. Electro-polished stainless steel tubes and metal sealedvalves were used. The gas cell and gas line were heated to200 ◦ C and pumped by a turbo molecular pump backed bya scroll pump. The argon gas was regulated using a needlevalve and injected into the inlet of the gas cell via a purifier(SAES Pure Gas, Inc FT400-902).
Fig. 3. The shape of the gas inlet was carefully fabricated by refer-encing flow simulation results. The gas is isotropically spread in thegas cell.
The differential pumping system
The differential pumping system is key to realize the cou-pling of gas cell and isotope separator. Historically, it hasbeen used in the IGISOL technique [15], where the pump-ing capability is a critical parameter to determine the oper-ational pressure for buffer stopping gas and the size of thegas cell exit hole. High pressures required for stopping en-ergetic RI-ions efficiently and a large exit hole for fast evac-uation to avoid the decay losses in short-lived RIs make fora severe pumping load in the differential pumping system.A typical IGISOL differential pumping system is shown inFig. 4 (top).When a compressible viscous gas flow passes through anorifice separating different pressures regions, the gas flowvelocity becomes supersonic at a certain pressure difference.Then the flow rate Q becomes independent of the pressurein the expansion area. The expression for Q in this case isgiven by [16] Q = AP (cid:18) γ + 1 (cid:19) γ − (cid:18) R T M γγ + 1 (cid:19) , (1)where A is the cross section of the exit hole (cm ), P thepressure in the gas cell, γ the ratio of specific heats C p / C v , R the gas constant, M the atomic mass of the gas and T the gas temperature.Additionally, some relationships relating flow rate topumping speed and conductance can be described in thissystem as follows :3 ig. 4. Typical IGISOL layouts of the gas cell and its connection tothe extraction chamber of the isotope separator (top) and conceptuallayout newly proposed differential pumping system connecting thegas cell to the mass separator (bottom). Q = Q jet + ( P − P ) C + P S , (2) Q jet + ( P − P ) C = ( P − P a ) C + P S , (3)( P − P a ) C = P a S a , (4)where Q jet is the fraction of the flow rate comprised bythe jet component which is directed into the SPIG. P , P and P a are the pressures of the individual rooms: P is thefirst room, containing the gas cell in the housing chamberon the beam line from accelerator; P is the middle room inthe differential pumping region; and P a is the third room,labelled as acceleration chamber in Fig. 4. C and C arethe conductances from the orifices connected between thefirst to the middle and the middle to the third room. S , S and S a are the effective pumping speeds connected withindividual room.For proper ion acceleration without discharge in the sep-arator, P a should be on the order of 10 − Pa or less. There-fore, it is necessary for P be on the order of 10 − Pa. Thiscan be easily deduced from eq. (4) by assuming the typicalconductance of extractor electrode to be C = 10 − m /sand standard pumping speed (a few m /s) of turbo molec-ular or diffusion pumps for S and S a . Consequently thelast term in the right-hand side of eq. (2), P S , should benearly equal to the total flow rate Q . This means that mostof the gas flow emitted from the gas cell is evacuated by thefirst pump S ,resulting in a severe pumping load. For in-stance if we take the following parameters: S = 2.0 m /s, S a = 1.0 m /s, P = 1.0 Pa, P = 1.0 × − Pa and P =1.0 × − Pa, then pumping speed S must be nearly 6 × m /h in order to use exit hole diameter of 0.5 mm whilekeeping a gas cell pressure of a few hundred hPa argon.In the present idea, however, the differential pumpingrooms up to the acceleration chamber are divided into morethan three separately pumped volumes, as shown in Fig. 4(bottom). Instead of immediately removing the buffer gasby a high-throughput pump, the pressure is stepwise de- creased across several rooms by means of small pumps. Ac-cordingly, the total flow rate Q can be also written as: Q = R X i =1 P i S i + P a S a . (5)where R is the number of differential pumping rooms.Thus the majority of the flow rate Q is divided by a num-ber of evacuation rooms to achieve a high-vacuum. Whilethe pumping speed S a required to achieve high-vacuum inthe acceleration chamber still needs high-throughput turbomolecular pumps, S i can be achieved with small pumps,resulting in a very compact differential pumping system.Based on this idea, we developed a differential pump-ing system as shown in Fig. 2. There are three differentialpumping rooms before high vacuum. The first and the sec-ond rooms are in built using a reentrant NW 50 chambermounted inside the third room. The first evacuation roomhas two outlet flanges, one is connected to a scroll pump(ANEST IWATA, 500 L/m), and the other to a pressuregauge. The second evacuation room is evacuated by a dryscrew pump (PFEIFFER Ontool booster pump, 36 L/s),while the third chamber is pumped by a turbo molecularpump (PFEIFFER HiPace300, 300L/s).2.3. The SextuPole Ion beam Guide (SPIG)
A 253 mm long SPIG is mounted parallel to the beampath between the exit hole and the Quadrupole Mass Sep-arator (QMS). The SPIG is built from six 1 mm diameterMolybdenum rods. The distance between the SPIG and thegas cell exit hole is 1 mm and the diameter of the inner cir-cle is 2 mm. A pair of RF signals are applied to the SPIGsuch that adjacent rods receive signals 180 ◦ out of phase.The frequency is fixed at 4.0 MHz and the amplitude canbe varied from 0 to 200 Vpp. A DC offset voltage can beadded to the RF signals.One concern is whether ions can be transported by SPIGdespite the long path (253 mm) in a large pressure gradi-ent. One may expect that all ions are simply stopped bycollisions with buffer gas before they reach the high vac-uum region. In the present case, however, a gas jet assistsions moving inside the SPIG from the high pressure regionof the exit hole to the high-vacuum region at the entranceof QMS.2.4. The laser optical system
A laser system for the resonant ionization in off-line setupis shown in Fig. 5. Two XeCl excimer lasers with wave-length λ = 308 nm (Lambda Physik LPX240i) pump twodye lasers (Lambda Physik Scanmate, FL3002). Both ex-cimer lasers are synchronized within a precision of a fewns. The maximum repetition rate of the excimer lasers is400 Hz. We use two-step one- or two-colour schemes for theresonant laser ionization of atoms. The first laser excites4toms from the ground state into an intermediate state fol-lowed by a transition into an auto-ionizing state or into theionization continuum by the second laser. A UV radiationis used for the first step; it can be synthesized by a secondharmonic generator (Lambda Physik UV311). A PC soft-ware package has been developed for remotely controllingall laser parameters, adjustment of timing delay, tuning,and scanning of wavelength. In the present prototype sys-tem, the laser beams are split into two paths, one path leadsto a reference cell where resonance ionization takes placein vacuum for testing the ionization scheme, the other pathleads to the laser ionization gas cell. Fig. 5. Scheme of the laser optics system.
3. Experimental results and discussion
Pressure distribution and gas jet behavior
The pressure at the individual evacuation rooms weremeasured by appropriate pressure gauges as shown inFig. 6. As described in previous section, in this new dif-ferential pumping method, the pressure is gradually de-creasing from the gas cell to the QMS. Even with nearlyone atmospheric gas cell pressure and 1 mm of exit holediameter, this differential pumping method is still capableof achieving high vacuum.By using Eq. (2) and present measured values, it is pos-sible to see the relative evolution of the jet component Q jet which is directly injected into the SPIG. As Q jet is a partialflow rate of Q , it is given by Q jet = εQ, (6)Then ε can be written from the relation in Eq. (2) asfollows: ε = 1 − ( P − P ) C + P S Q . (7)In the intermediate vacuum region, the conductancevaries with the pressure slightly. If we use a fixed value of C = 3.8 × − m /s which is calculated for the mechan-ical structure, and substituted Q from Eq. (1) and S =0.5 m /min, then the behavior of ε can be approximatedby using experimentally measured pressures P and P ,as shown in Fig. 7. Fig. 6. Measured pressures at gas cell, each evacuation room in thedifferential pumping region and QMS. The size of gas cell exit holediameter is 1 mm.Fig. 7. Deduced gas jet fraction ε using measured pressures as func-tion of gas cell pressure. Absolute values of ε cannot be reliably determined, sincethe pressure gauges are placed at the wall of chamber inthe evacuation rooms, these values differ from that of realbeam path. However it can be presumed that the gas jetcomponent directed into the SPIG increases with the back-ground pressure. This means that the divergence of gas jetdecreases as the pressure of the first differential pumpingregion increases [17]. This partial flow rate is regarded asan assisting force gas jet which brings ions to high vacuumduring the SPIG transmission.3.2. Resonant ionization inside the gas cell, ion extractionand transport to high vacuum
The resonant laser ionization inside the gas cell, ion ex-traction and transport to the high-vacuum region via SPIG5nd QMS have been confirmed. So far we have tested ion-ization inside the gas cell for stable isotopes of the follow-ing elements: Ni, Fe, Cu, Co, Pd, Sn, Ti, Nb and In. Theionization was performed by one colour or two colours two-step excitation scheme [19].
Fig. 8. Scans of the first step laser wavelength for resonance ionizationof Fe, Co and Nb. The second step laser was tuned to wavelengthsfor the auto-ionizing state in case of Fe and Co.
Atoms to be photo-ionized were produced by either evap-oration of a filament or YAG laser ablation inside the cell.The gas cell pressure was adjusted from 100 to 700 hPaargon. First, evaporated atoms are released in the middleregion of the gas cell, then they move via gas flow to thelaser ionization region just before the gas cell exit. Two dyelaser beams are focused on a quartz view port and irradiatethe flowing atoms for ionization. The laser repetition ratewas typically set between 10-50 Hz during the experiment,though this value can be increased to a maximum of 400Hz.Since the laser wavelength is tuned for an element-specific ionization scheme which is determined by anionizing atomic transition, we can identify the ion of in-terest with an element separation by laser in addition to amass separation by QMS. After ionization inside the gascell, photo-ionized ions are injected into the SPIG witha gas jet, and pass through the QMS entrance with 10mm aperture, providing mass selection prior to CEM. Fig- ure 8 shows example detected ion signals for Fe, Co andNb in scans of the first step excitation laser wavelength.Ion detection rate depends on the status of filament orablation target; however, normally more than 10 countsper second (cps) are observed with 50 Hz laser repetitionrate. Saturation of ionization efficiency was observed ineach element, typically around 100 µ J/pulse in the firsttransition, 1 mJ/pulse in the auto-ionizing transition.3.3.
The investigation of SPIG performance
The investigation of SPIG performance was carried outby both simulation and experiment. In simulation, the tra-jectories for individual ions were microscopically tracedstep by step using a Monte Carlo technique until the ionseither collide with SPIG rods or reach the end of the SPIG.The effects of the RF + DC electric fields and gas flow af-fecting the motion of ions were calculated separately. TheNavier-Stokes flow calculation provides flow velocity, tem-perature and density for the gas. The calculation used amesh with dimensions of 10-100 µ m, limited by the com-puting memory and reasonable time of computation. Themean free path of ions was determined by the density, whilethe velocity of the buffer gas atoms was calculated by avector summation of the flow velocity and the thermal mo-tion of gas atoms following the Maxwell-Boltzmann distri-bution. Electric field was calculated by SIMION [18]. Be-tween collisions, the motion of the ions under the influenceof the RF+DC electric fields was determined by means of a4 th order Runge-Kutta method. Collisions kinematics werecalculated by the rigid sphere model. While there is noway to exactly reproduce the experimental result by sim-ulation without knowing the real pressure distribution in-side the SPIG, it is possible to understand the behavior ofion transmission by calculating pressure distribution usingmeasured pressures as boundary conditions. In this work,the tested boundary conditions were as follows: 500 hPa inthe gas cell, 10 hPa in the first differential pumping room,1 hPa in the second differential pumping room, and 10 − hPa in the third differential pumping room. A continuouspressure gradient, based on these boundary conditions, wasutilized in simulations.In this simulation, the initial ion distribution was ran-domly populated, using a guassian random number gener-ator by width of 0.5 mm, on a surface filling the opening ofthe exit hole. Space charge effects and ion loss due to inter-action between impurities inside buffer gas were not con-sidered. A typical trajectory for an Fe + ion moving throughthe SPIG in the simulation is shown in Fig. 9; the simula-tion used SPIG RF signals of 100 V pp at 4.0 MHz. From thisfigure, we can see that the ion can be transmitted throughthe SPIG, even in the case of relatively dense pressure gra-dient.The simulated transmission efficiency as a function of RFvoltage is shown in Fig. 10 for the pressure distribution de-scribed above. The graph also includes experimental data,6 ig. 9. Monte Carlo simulation of an ion’s motion in the SPIG. Anion ( m = 56) starting at the gas cell exit hole follows the gas flow,along the z axis, of the assisting force gas jet while trapped in theradial potential well of the SPIG RF field. measured with 500 hPa argon in the gas cell. The simu-lated efficiency scale is given along the left vertical axis,while the right vertical axis denotes experimentally mea-sured ion intensity. The simulated efficiency is based on ini-tial ion distribution of 100 ions with atomic mass of m =56. The experimental data was measured by operating theQMS with RF and DC potentials appropriate for rejectingall ions with m/q = 56 without DC potential difference be-tween SPIG and QMS. The trend in the experiment andcalculated curves agree well. A saturation voltage, wherea radial focusing force is enough to confine ions inside theSPIG, was found near 80 V pp , in good agreement with thesimulation.Separately, it was possible to determine the absolutetransmission efficiency for the SPIG. The experimental se-tups and the used ionization scheme for Cu are shown inFig. 11. The efficiency is defined as the ratio of the ioncurrent operated with RF signals of 100 V pp at 4.0 MHzcollected on a Faraday cup placed after the SPIG (SET Bin Fig. 11) to the ion current collected on the SPIG rodswithout RF signal applied (SET A in Fig. 11). This wasdone during a test of resonant laser ionization for stableCu. Comparing from the integrated value of two ionizationcurves, shown in Fig. 12, an absolute SPIG transimissionefficiency was 60-80% with a gas cell pressure of 500 hPa.3.4. Time profiles of photo-ionized ions as a probe of SPIGtransit time and gas cell extraction time
The time profile of ions extracted from the gas cell pro-vides an important probe to understand the behavior of
Fig. 10. Simulated and measured SPIG transmission of Fe + as func-tions of RF voltage applied to SPIG. Simulations followed ions to theend of the SPIG, while measured data required ions to pass througha QMS set to select for mq = 56Fig. 11. Layout of the apparatus used for the measurement of theSPIG transmission efficiency. SET A was used for the measurementof the total ion current extracted from the gas cell. The ions werecollected on on the SPIG rods, without RF signal. While SET Bwas used to measure the ion current transmitted to the end of SPIGoperated with RF signals of 100 V pp at 4.0 MHz. Atomic leveldiagram of Cu indicated on the left side is the ionization schemeused in the present experiment. ions moving inside the gas cell and the SPIG. In this work,we have examined the time profile in two trigger modes. Us-ing the ionization laser pulse as a start signal probes trans-mission time through the SPIG, while using a YAG laserpulse as a start signal allows the use of ablated material asa probe of extraction time from the gas cell. Ion detectionat the CEM provides the stop signal in both cases.3.4.1. Transport of ions through the SPIG
It is necessary to know the time required for ion trans-port through the differential pumping region. Due to thehigh pressure distribution in the SPIG, it is not obviousthat the transport will be sufficiently fast. To investigatethis effect, the ionization lasers were operated at 10 Hzand serve as a start signal for the timing measurement.Fe vapor atoms were continuously released from a hot fila-7 ig. 12. Ion current as function of first-step laser wavelength, mea-sured during laser resonance ionization of Cu. The dashed line iscurrent measured on the SPIG rods, without RF signal applied (SETA). The solid is current measured at a Faraday cup immediately fol-lowing the SPIG, when the SPIG was operated with RF signals of100 V pp at 4.0 MHz (SET B). From the integrated value of the twocurves, the absolute SPIG transimission efficiency was 60-80%. ment inside the gas cell. They were resonantly ionized andfinally measured at the CEM behind the QMS, with theQMS set to selectively transport ions with m/q = 56. Thetime profile results for Fe + using various gas cell pressuresare shown in Fig. 13. The signal represents a convolutionof three components: transport from the ionization regionto the gas cell exit hole, transport through the SPIG andtransport through the QMS. We can neglect the transporttime through the QMS, as it is located in high-vacuum andtransport should occur in the order of tens of microseconds. Fig. 13. The measured extraction time profile from the ionizationregion to the CEM ion detector.
These results show that the transport time for ions be-comes faster with increasing gas cell pressure. For the firstorder, the extraction time in the gas cell does not dependon the pressure, implying that this phenomena is caused bythe reduction of the SPIG transmission time. This can beunderstood from the shape of gas jet. When the pressure inthe gas cell increases, the background pressure in the firstdifferential pumping room is also increased and the diver-gence angle of the gas jet becomes small as described in section 3.1 and [17]. This causes the effective length of gasjet inside the SPIG to increase. As a result, successive col-lisions force ions towards the high-vacuum region. This re-duces transmission time for ions in the SPIG. On the otherhand, once the gas jet disappears, ions are no longer effec-tively pushed, which leads to longer transmission times atlower gas cell pressures.
Fig. 14. Simulated gas velocity distribution at different pressureconditions in individual room. The colour scale covers the range 0to 300 m/s . Colour online.
Figure 14 shows the flow simulation result revealing thegas velocity distribution for various gas cell pressures atcutting plane of the central axis of the SPIG. The boundaryconditions for gas cell pressure used in the simulation areshown at the left side in the figure. It can be seen that thesimulated gas jet length in terms of the gas velocity over300 m / s (red colour) becomes longer as the gas cell pressureincreases. This gas jet length indicates the effective rangeof the assisting force gas jet, such the higher velocity gasjet brings ions to the vacuum region quickly.We can conclude from the qualitative agreement betweensimulation and experiment that the jet structure deter-mines the transit time of ions inside the SPIG. This assist-ing force gas jet effect can be utilized to provide effectivetransmission of ions into the high-vacuum region despitethe high background pressure, extending the permissibleupper pressure limit in the gas cell.3.4.2. Extraction of atoms from the gas cell by gas flow
A laser ablation target was installed in the gas cell, ap-proximately 60 mm from the gas cell exit nozzle. This al-lows for the pulsed production of atomic ensemble by YAGlaser. By operating the YAG laser at a repetition rate of 18z, while the repetition rate of ionization lasers is 100 Hz,it is possible to study the transport time of atoms in thegas cell. Figure 15 compares simulated and measured timeprofiles for stable Fe, using the ablation pulse of the YAGlaser as the start signal, with gas cell pressure of 100 hPa.When the wavelength of the ionization laser was off-resonance, no ion signal was observed with QMS selectivelytransmitting ions with m/q = 56. This indicates that mostFe + ions produced by YAG laser ablation are neutralized inthe gas cell. The time-scale is an order of magnitude largerthan that for SPIG transmission time shown in Fig. 13.Most of the extraction time is spent transporting ions froma region near the ablation target to the exit of the cell. Thebroad time-structure in Fig. 15 can be explained by theinitial distribution of atoms in the ablation plume. Fig. 15. The time profile for atom’s released by YAG laser ablationin the gas cell, photo-ionized at the exit nozzle and measured atCEM. Good agreement can be seen between measured (points) andsimulated (line) results.
The simulated results were calculated using macroscopicmotion combined with gas flow calculations. The gas flowcalculation was performed using a precise 3D model of thegas cell which took the full geometrical complexity intoaccount. The gas flow calculation provided flow velocities ateach point on a 3D mesh. A random number generator wasused to create an initial distribution of the atomic plumein a region near the ablation target surface. Trajectorieswere then calculated from the gas flow, in combination withdiffusion of the atomic plume, by means of a fixed time-step ray-tracing algorithm until collision with the chamberwall or transport to the exit region. Good agreement is seenbetween the two results.The experimental data include transport time throughthe SPIG and QMS, while the simulation does not includethese transmission times as it ends when ions reach the gascell exit hole. This makes for a discrepancy between ex-periment and simulation, because the rise time from sim-ulations is artificially fast from lacking transmission timestrough the SPIG and QMS.
4. Summary and outlook
We have shown proof of principle studies for a resonantlaser ionization gas cell that utilizes a novel differentialpumping method, based on the use of a few small scrollpumps. In terms of maximum allowable gas cell pressures,the new method was shown to be comparable or superior toconventional differential pumping systems based on largeroots or screw pumps configuration. Furthermore the sizeof the differential pumping system has been drastically re-duced. It will open a new possibility for breaking throughthe limitations on gas cell pressure and exit hole size, whileallowing for a very compact assembly.Off-line tests of resonant laser ionization for stable Co,Cu, Fe, Ni, Ti, Pd, Nb, Sn and In inside the gas cell, alongwith extraction of these ions to high-vacuum, was success-ful. The ability to photo-ionize atomic vapors and trans-port the ions through a long SPIG was confirmed. The fastion transport can be attributed to the affect of an assistingforce gas jet. This gas jet is an effect of allowing a high pres-sure in the first differential pumping room, which results ina narrow, collimated gas jet. This narrow, collimated gas jetforces ions through the SPIG quickly. As another spin-off,this jet structure is brought for ideal environment for laserspectroscopy[20][21][22][23]. The first result for the funda-mental study of gas-jet laser spectroscopy in the presentsetup can be found in [24].Simulation of atomic transport in the gas cell and iontransport through the SPIG was experimentally verified.By using the ionizing laser pulse as a start signal, it waspossible to study the transit time through the SPIG. Theresult showed an inverse relationship between gas cell pres-sure and SPIG transit time. This is a powerful verificationof the existence of an assisting force gas jet. Alternately, theuse of a YAG laser to produce pulsed atomic vapor plumesallowed the study of transport inside the gas cell. The ex-cellent agreement of these measurements with simulationsbased on gas flow calculations allows us to confidently es-timate the expected yield of short-lived nuclei for newlydesigned PALIS gas cell.The miniaturization of the differential pumping system,along with the ability to use higher gas pressures, will en-able installation of the parasitic laser ion source (PALIS)in the limited space of the BigRIPS fragment separator.As the first phase, PALIS will be installed in the vicin-ity of F2 focal plane in the BigRIPS fragment separator.Detailed design work for the gas cell, differential pump-ing system and the low energy RI-beam transport line iscurrently in progress. The expected PALIS yield, based onextraction time from gas cell and ability to photo-ionize,is shown in Fig. 16. The availability of PALIS system willresult in a many-fold increase in the available beam time.Furthermore, in combination with narrow gas jet structure,the laser spectroscopy inside the gas jet will be a vital toolto explore the study of the ground state property for widerange of rare exotic nuclei provided by BigRIPS.9 ig. 16. The expected yield of available elements by laser ionizationfor PALIS on-line experiments. The estimation includes efficienciesattributed to individual processes, such as stopping in the gas cell,laser ionization and decay loss for short-lived nuclei.
5. Acknowledgments
We wish to thank RIKEN Nishina Center for Accelerator-Based Science for financial support of this research. Wewould like to thank Dr. M. Wakasugi of RIKEN, Dr.M. Koizumi and Dr. M. Ohba of JAEA and Dr. T. Mitsug-ashira of IRCNMS for providing us their laser components.Also we would like to express our appreciation of Mr.H. Imamura of Hakuto Co.,Ltd, Mr. T. Kimura of CoherentJapan, Ms. W. Kina of Indeco, INC. and Dr. T. Kambaraand Dr. T. Kobayashi of RIKEN for their kind support. Weare grateful to M. Huyse, P. Van Duppen, Yu. Kudryavt-sev, R. Ferrer and P. Van den Bergh of K.U.Leuven fortheir support and fruitful discussions during development.
References [1] M. Wada, Y. Ishida, T. Nakamura, Y. Yamazaki, T. Kambara,H. Ohyama, Y. Kanai, T.M. Kojima, Y. Nakai, N. Ohshima,A. Yoshida, T. Kubo,Y. Matsuo, Y. Fukuyama, K. Okada,T. Sonoda, S. Ohtani, K. Noda, H. Kawakami, and I. Katayama,Nucl. Instr. Meth. B (2003) 570.[2] M. Wada, A. Takamine, T. Sonoda, P. Schury, K. Okada,SLOWRI Collaboration, Hyp. Int. (2010) 43.[3] T. Sonoda, M. Wada, A. Takamine, K. Okada, P. Schury, A.Yoshida, T. Kubo, Y. Matsuo, T. Furukawa, T. Wakui, T.Shinozuka, H. Iimura, Y. Yamazaki, I. Katayama, S. Ohtani,H. Wollnik, H.A. Schuessler, Y. Kudryatsev, P. Van Duppen,M. Huyse, the SLOWRI Collaboration: AIP Conf. Proc. (2009) 132.[4] T. Kubo, Nucl. Instr. Meth. B204 (2003) 97.[5] G. Savard, J. Schwartz, J. Caggiano, J.P. Greene, A. Heinz, M.Maier, D. Seweryniak, B.J. Zabransky, Nucl. Phys. A (2002)292.[6] L. Weissman, P.A. Lofy, D.A. Davies, D.J. Morrissey, P. Schury,S. Schwarz, T. Sun, G. Bollen, Nucl. Phys. A (2004) 655.[7] L. Vermeeren, N. Bijnens, M. Huyse, Y. A. Kudryavtsev, P. VanDuppen, J. Wauters, Z. N. Qamhieh, P. Thoen, E. Vandeweert,R. E. Silverans, Phys. Rev. Lett. (1994) 1935.[8] P. Van Duppen, Nucl. Inst. and Meth. B (1997) 66.[9] Yu. Kudryavtsev, J. Andrzejewski, N. Bijnens, S. Franchoo,J. Gentens, M. Huyse, A. Piechaczek, J. Szerypo, I. Reusen,P. Van Duppen, P. Van den Bergh, L. Vermeeren, J. Wauters,A. W¨ohr, Nucl. Inst. and Meth. B (1996) 350. [10] Yu. Kudryavtsev, T.E. Cocolios, J. Gentens, M. Huyse,O. Ivanov, D. Pauwels, T. Sonoda, P. Van den Bergh, P. VanDuppen, Nucl. Inst. and Meth. B (2009) 2908.[11] H.J. Xu, M. Wada, J. Tanaka, H. Kawakami, S. Ohtani andI. Katayama, Nucl. Inst. and Meth. A (1993) 274.[12] S. Fujitaka, M. Wada, H. Wang, J. Tanaka, H. Kawakami,I. Katayama, K. Ogino, H. Katsuragawa, T. Nakamura,K. Okada, S. Ohtani, Nucl. Inst. and Meth. B (1997) 386.[13] P. Van den Bergh, S. Franchoo, J. Gentens, M. Huyse,Yu. Kudryavtsev, A. Piechaczek, R. Raabe, I. Reusen, P. VanDuppen, L. Vermeeren, A. W¨ohr, Nucl. Inst. and Meth. B (1997) 194.[14] P. Karvonen, I.D. Moore, T. Sonoda, T. Kessler, H. Penttil¨a,K. Per¨aj¨arvi, P. Ronkanen, J. ¨Ayst¨o, Nucl. Inst. and Meth. B (2008) 4794.[15] J. ¨Ayst¨o, Nucl. Phys A (2001) 477.[16] A. Roth, Vacuum Technology (North-Holland, Amsterdam,1982).[17] M. Reponen, I.D Moore, I. Pohjalainen, T. Kessler, P. Karvonen,J. Kurpeta, B. Marsh, S. Piszczek, V. Sonnenschein and J. ¨Ayst¨o,Nucl. Inst. and Meth. A (2011) 24.[18] SIMION 3D Version 7.0 Ion Optics Program, 2006.[19] V.N. Fedosseev, Yu. Kudryavtsev, V.I. Mishin, Phys. Scr. (2012) 058104.[20] R. Ferrer-Garcia, V. T. Sonnenschein, B. Bastin, S. Franchoo,M. Huyse, Yu. Kudryavtsev, T. Kron, N. Lecesne, I. D. Moore,B. Osmond, D. Pauwels, D. Radulov, S. Reader,1, L. Rens, M.Reponen, J. Rossnagel, H. Savajols, T. Sonoda, J. C. Thomas,P. Van den Bergh, P. Van Duppen, K. Wendt, S. Zemlyanoy,submitted to Nucl. Inst. Meth. B.[21] T. Sonoda, T.E. Cocolios, J. Gentens, M. Huyse, O. Ivanov, Yu.Kudryavtsev, D. Pauwels, P. Van den Bergh, P. Van Duppen,Nucl. Inst. Meth. B (2009) 2918.[22] I.D. Moore, A. Nieminen, J. Billowes, P. Campbell, Ch. Geppert,A. Jokinen, T. Kessler, B. Marsh, H. Penttila, S. Rinta-Antila,B. Tordoff, K. Wendt and J. ¨Ayst¨o, J. Phys. G (2005) s1499.[23] P. Karvonen, T. Sonoda, I.D. Moore, J. Billowes, A. Jokinen,T. Kessler, H. Penttila, A. Popov, B. Tordoff and J. ¨Ayst¨o,Eur. Phys. J. Special Topics (2007) 283.[24] T. Takatsuka, H. Tomita, T. Sonoda, V. Sonnenschein,C. Sakamoto, H. Mita, T. Noto, C. Ito, S. Maeda, T. Iguchi,M. Wada, K. Wendt, I.D. Moore, to be published in Hyp. Int.(LAP 2012 proceedings).(2007) 283.[24] T. Takatsuka, H. Tomita, T. Sonoda, V. Sonnenschein,C. Sakamoto, H. Mita, T. Noto, C. Ito, S. Maeda, T. Iguchi,M. Wada, K. Wendt, I.D. Moore, to be published in Hyp. Int.(LAP 2012 proceedings).