- Email: [email protected]
A new fast-cycling system for AMS at ANU
Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Contents lists available at ScienceDirect
Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb
A new fast-cycling system for AMS at ANU M. De Cesare ⇑, L.K. Fifield, D.C. Weisser, D. Tsifakis, A. Cooper, N.R. Lobanov, T.B. Tunningley, S.G. Tims, A. Wallner Department of Nuclear Physics, Research School of Physics and Engineering, Australian National University, ACT 2601, Canberra, Australia
a r t i c l e
i n f o
Article history: Received 15 December 2014 Received in revised form 20 May 2015 Accepted 20 May 2015 Available online 29 May 2015 Keywords: Accelerator mass spectrometry Fast-cycling system Actinides
a b s t r a c t In order to perform higher precision measurements, an upgrade of the ANU accelerator is underway. Fast switching times on the low-energy side, with maximum settling times of 30 ms, are achieved by holding the injector magnet field constant while changing the energy of the different isotopes by changing the pre-acceleration voltage after the ion source. Because ions of the different isotopes then have different energies before injection, it is necessary to adjust the strength and steering of the electrostatic quadrupole lens that focusses the beam before entry into the accelerator. First tests of the low-energy system will be reported. At the high energy end, a larger vacuum box in the analyzing magnet has been designed, manufactured and installed to allow the transport of differences in mass as large as 10% at constant terminal voltage. For the cases where more than one isotope must be transported to the detector an additional refinement is necessary. If the accelerator voltage is to be kept constant, then the trajectories of the different isotopes around both the analyzing and switching magnets must be modified. This will be achieved using bounced electrostatic steerers before and after the magnets. Simulations have been performed with the ion optic code COSY Infinity to determine the optimal positions and sizes of these steerers. Ó 2015 Published by Elsevier B.V.
1. Introduction Due to the exigencies of sharing a large accelerator with an active nuclear physics program, and limitations imposed by existing hardware, AMS at the Australian National University [1] (ANU) has for many years operated in a sequential slow-cycling mode. Counting times for the rare isotope are typically two to five minutes, and are dictated by the requirement to maximize the fraction of time spent on counting the rare isotope(s) when switching times between isotopes are 10–15 s. Ion source output can, however, vary appreciably during these counting times, which limits the precision that can presently be achieved to 3%. In order to perform higher precision measurements, an upgrade of the ANU accelerator is underway to allow a fast-cycling procedure. Here we describe the various components of the upgrade and its present status. 2. Upgrade of the ANU accelerator and actinide measurements The ANU AMS system is based on a 15 MV tandem accelerator. As high energy is required to apply certain techniques of isobar ⇑ Corresponding author at: Italian Aerospace Research Centre – CIRA, Via Maiorise, 81043 Capua (CE), Italy. E-mail address: [email protected] (M. De Cesare). http://dx.doi.org/10.1016/j.nimb.2015.05.028 0168-583X/Ó 2015 Published by Elsevier B.V.
separation effectively, this makes the ANU tandem particularly well-suited for the heavier isotopes with severe isobaric interferences e.g., 36Cl and 53Mn, but it is sufficiently flexible that it can also be run at lower voltages 4 MV for actinide measurements [1,2]. A schematic layout of the ANU 15 MV tandem facility is shown in Fig. 1. The caesium sputter ion source is a 32-sample MC-SNICS. Negative ions from this source are pre-accelerated and mass-analysed by the mass-energy product 56 MeV amu/e2 injection magnet, which allows high resolution mass analysis for all stable isotopes in the periodic table. The injection beam line also features a fast (rise-time 50 ns) electrostatic chopper that allows attenuation of beam currents that would be too high for injection into the tandem accelerator (e.g. 35Cl) or counting rates that would be too high for the detector (e.g. 234U). An electrostatic quadrupole triplet with steering capability focusses the beams to a waist inside the accelerator tank. This waist is the object point for the lens formed by the electric field at the entrance to the acceleration tube which then focusses the beams into the stripper canal in the high voltage terminal. The 14UD accelerator is capable of voltages up to 15 MV and is routinely operated above 14 MV. Both a gas stripper and a foil stripper are available at the terminal. The double focusing high-energy analyzing magnet has a radius of 1.27 m and a mass-energy product of 225 MeV amu/e2 at its maximum field of 1.7 T. Its vacuum box has an opening in the bending direction of
476
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Fig. 1. Schematic layout of the ANU 14UD accelerator and the 15° beam line that is used inter alia for AMS of actinides showing actual and planned upgrades for fast-cycling AMS operation. In the actinides beam line A denotes a selectable 3 or 6 mm diameter aperture. The accelerator and analysing magnet are vertical, while the switching magnet bends in the horizontal plane. The upgraded part includes the ±10 kV TREK power supply in the pre-acceleration to switch between different isotopes, the new analyzing magnet chamber and the offset Faraday cups after the analyzing magnet. The deflector plates before and after the analyzing and switching magnets will be installed at a later date. Note that the ionization chamber is removed when either the time-of-flight system or the gas-filled magnet are being used.
only 59 mm at the exit from the magnet. This is insufficient to allow 35Cl, for example, to be transmitted at the same time as 36 Cl when the latter is on the central axis, and has been a major inhibiting factor to the development of a fast-cycling capability. Two beam lines are available for AMS, one equipped with a multi-element gas ionization detector, the other with a 6 m time-of-flight system, a gas-filled magnet, and an ionization
detector. Wien filters on both beam lines provide the final analysis stage to remove ions with the same mE/q2 but different E/q. In the slow-cycling mode, different isotopes are injected into the accelerator by changing the magnetic field in the injection magnet. On the high-energy side, the analyzing and switching magnets are held fixed while the terminal voltage as well as the electric field in the Wien filter are changed in order to transmit
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
all the isotopes to the detector station. For example, 35Cl, 37Cl and 36 Cl are injected in a cycle in which the beam currents of 35Cl and 37 Cl are measured in a Faraday cup immediately in front of the gas ionization detector for 10 s each, while the 36Cl ions are counted in the detector for 5 min. In the case of plutonium, 242Pu, 240Pu and 239 Pu are all counted in the detector for typically 1, 3 and 2 min respectively. For the fast-cycling system presently being implemented, fast switching times on the low-energy side will be achieved by holding the injector magnet field constant while changing the energy of the different isotopes by changing the pre-acceleration voltage after the ion source. This normally operates at 150 kV. The incremental voltages required for fast switching between isotopes will be supplied by a fast high-voltage amplifier (±10 kV from TREK) in series with the Glassman 200 kV supply [3]. Because the energy of the injected beam is changed, it is also necessary to adjust the strength of an electrostatic quadrupole triplet before injection, and this will be achieved with six additional TREK high-voltage amplifiers. At the high energy end a larger vacuum box in the analyzing magnet has been designed, manufactured and installed to allow the transport of isotopes differing in mass by up to 10% (e.g. 9Be and 10Be). Currents of the stable beams (e.g. 35Cl and 37 Cl) will be measured in offset Faraday cups after the analyzing magnet, and the appropriate vacuum housing and Faraday cups have been designed and constructed on the basis of detailed beam optics calculations using the code COSY Infinity [4]. These cups are attached to motorized linear drives so that they can be positioned appropriately for the wide range of isotopes measured with the 14UD accelerator. For the cases where more than one isotope must be transported to the detector (e.g. 239,240,242Pu or 233,236U) an additional refinement is necessary. If the accelerator voltage is to be kept constant, then the trajectories of the different isotopes around both the analyzing and switching magnets must be modified. This will be achieved using bounced electrostatic steerers before and after the magnets, as implemented for example at ANSTO for actinide measurements and formerly at the Sydney CSIRO facility for ‘super-SIMS’ applications [5,6]. Simulations have been performed with COSY Infinity to determine the optimal positions and sizes of these steerers.
477
160 kV to 140 kV. Clearly, the response time is long compared to the 100 ls rise time of the TREK amplifier due to the time taken for the Glassman supply to adjust to the changed current demand. The manufacturer quotes a time constant of 50 ms for the response of the Glassman supply to a demand for a voltage change, so it is not surprising that the response time of the voltage stabilisation circuit to a change in current is of the same order. Of course, this 20 kV swing is the extreme case, and actual changes will be substantially smaller. Switching between 35Cl and 37Cl, for example, will require a change from about 4 kV to +4 kV where the response is faster. Nevertheless, it is necessary to wait for 30 ms for the accelerating voltage to stabilise before performing the current measurements of the stable beams, or starting the counting of the rare isotope. During this 30 ms, the low-energy chopper will be employed as a ‘blanking steerer’ to sweep the beam onto the image slits of the injection magnet so that it is not injected into the accelerator. In order to explore whether the response time could be shortened, an additional current drain from the Glassman supply was effected by adding a 200 MO resistor chain from high voltage to ground. As can be seen in Fig. 2, this steepens the rising edge of the pulse, but it subsequently takes longer to return to the baseline so was not pursued further. 3.2. Low-energy focus control The quadrupole triplet lens that forms the final waist before the beam enters the low-energy accelerator tube is electrostatic. Hence, its strength for given applied voltages depends on ion energy. Since the various ions have different energies in this lens, it is necessary to adjust the applied voltages from one ion species to another. In [3] preliminary measurements with a DC beam of 28 Si at a terminal voltage of 12.7 MV confirmed that the transmission decreased when the injection voltage was reduced or increased by 4 kV from the nominal 150 kV. In both cases, full transmission could be restored by small changes in the strength of the lens only. Since the lens incorporates steering as well as focussing, however, it requires six high-voltage power supplies and adjustments to all six are required in order to change the strength of the lens. The response times of the six Glassman power supplies that previously supplied these voltages were too slow to be compatible with the fast-switching sequence, and have been
3. Simulations and experimental results 3.1. Low-energy side of the accelerator Fast switching between isotopes in modern AMS facilities is almost universally accomplished by modifying the beam energy though the 90° injection magnet by applying voltages of several kV to the insulated vacuum box of the magnet. This option was not practicable at the ANU, both because of the presence of the fast beam chopper immediately after the injection magnet, and because the magnet must be rotated between two ion sources. So instead, the novel alternative of pulsing the voltage on the high voltage ion source deck is being implemented using a TREK 1040 A, ±10 kV power supply connected in series with the Glassman 200 kV power supply that presently supplies the pre-acceleration voltage of typically 150 kV. The TREK power supply is controlled via an analog ±10 V signal which is transferred from ground potential via a fiber optic cable [3]. It will be controlled by a sequencer made by NEC. A series of tests were performed in order to quantify the effectiveness of the TREK power supply in rapidly changing the ion source deck voltage. The results are summarised in Fig. 2, which shows the response of the ion source deck when the TREK amplifier was switched from +10 kV to 10 kV in order to change the accelerating voltage from
Fig. 2. Oscillograms showing the response of the voltage of the ion source deck when the Glassman high voltage supply is set to 150 kV and the TREK supply is switched at time = 0 ms from +10 to 10 kV (the chassis of the TREK is connected to the deck), and from 10 to +10 kV at time = 50 ms. The trace with the faster rise time was obtained with an additional 200 MO load on the Glassman supply (see text). The width of the ‘measuring time’ has been exaggerated for clarity.
478
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
replaced by six TREK 10 kV fast amplifiers. These are unipolar, with four being negative and two positive. The focussing at the entrance to the low-energy acceleration tube of the 14UD accelerator is controlled by a novel tube-entrance lens [7] that uses a 150 kV power supply. Its strength for a given voltage also depends on the energy of the injected ions, and in principle it should also be adjusted from isotope to isotope. It was not practicable to change this voltage, but the changes in focussing may be compensated by moving the beam-waist closer to or further from the tube-entrance lens with the quadrupole triplet. The strengths of the quadrupole triplet therefore differ slightly from what would be required to focus all isotopes to the same waist position, and are set empirically to optimise transmission for each isotope. 3.3. High-energy side of the accelerator The new analyzing magnet chamber, Fig. 3, and the subsequent box to house the off-axis Faraday cups were designed on the basis of simulations with the code COSY Infinity. In particular, the chamber and Faraday cup box are designed to accommodate off-axis 9Be ions when 10Be of the same energy is on the central trajectory. Accordingly the exit aperture is 180.6 mm in the bending plane, which is a factor of three larger than the 59.0 mm of the original vacuum chamber. The inside gap of the vacuum chamber parallel to the magnetic field is 23.75 mm, and the aperture in the bending plane at the entrance is 59.0 mm. In order to define the dimensions of the offset Faraday cups, we have simulated the displacement between the commonly-used isotopes that are the most difficult to separate, i.e. between the abundant 238U and the rare 236U and between the abundant 56Fe, 57 Fe and the rare 60Fe. The distance between the exit from the magnet and the Faraday cups is constrained to 80 cm by the need to rotate the analyzing magnet between two target rooms, since a beam line valve is required to isolate the magnet vacuum from
the downstream system when the magnet is being rotated. The off-axis Faraday cups are therefore not at the image plane of the 127 cm radius magnet and it is important to know the beam dimensions at their actual positions. The results of the simulation obtained with the COSY Infinity beam optics program are shown in Fig. 4. At the object point of the analyzing magnet the beam is assumed to have a full width of ±1.5 mm and a maximum angular divergence of 3 mrad which is the worst case scenario obtained with uranium. It is clear that 238U5+ and 236U5+ ions are not sufficiently separated at 80 cm from the magnet, but would be well separated if an additional off-axis cup were placed at 2 m from the magnet. This would be possible, since even at 2 m from the magnet the 238U ions are only 2 cm off axis and hence could be transported through the beam line valves and other apertures. In the case of the Fe isotopes, 56Fe and 57Fe are seen to be well separated from each other and from 60Fe even at 80 cm from the magnet. The beam full widths are 12 mm and the separation between 56 Fe and 57Fe is 24 mm, and hence the individual isotopes are comfortably accommodated in Faraday cups with inner and outer diameters of about 20 and 22 mm respectively. Tests with a 35Cl beam indicate that the beam size is indeed comfortably within the cup. Note that the 12 mm full width of the 60Fe beam is that at 80 cm from the magnet, and assumes an angular spread of 3 mrad, which is probably a significant overestimate. At the image point of the magnet, 254 cm from the magnet exit, the beam full width is 3 mm. Plutonium isotopes, extracted as molecular ions (xPu16O ) [2], must all be counted in the ionization chamber. In order to transmit 239,240,242 Pu sequentially to the detector at a fixed terminal voltage, the trajectories around the analyzing and switching magnets must be modified with electrostatic steerers, the voltages of which can be switched from beam to beam. Provision has been made to house such a deflector after the analyzing magnet by incorporating a deflector chamber immediately after the magnet as shown in Fig. 3. Similar chambers are envisaged before the analyzing magnet, and before and after the switching magnet. A simulation of
Fig. 3. The new magnet chamber, the deflector chamber and the chamber that houses three offset movable Faraday cups for the abundant isotopes, see text. The blue central ray represents the rare selected isotope, the two red (upper) rays represent lighter isotopes, while the green (lower) ray represents a heavier isotope. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
479
Fig. 4. Beam positions and widths after the analyzing magnet at the indicated distances from the magnet exit. The upper plot shows uranium isotopes, while the lower plot shows iron isotopes. It is assumed that, at the object point of the analysing magnet, the beam is uniformly distributed over a full width of ±1.5 mm and an angular spread of 3 mrad.
the beam trajectories for xPu isotopes from the object point of the analyzing magnet to the Ionization Chamber (Fig. 1) in the actinides beam line has been performed with the COSY Infinity beam optics program. Fig. 5 shows the result for a 23.846 MeV 242Pu5+ beam (4 MV terminal voltage with an injection energy of 100 keV), in the bending plane, X, and vertical plane, Y. The deflector plates are positioned 42 cm before and after the analysing and switching magnet, and have a length of 12 cm and a gap of 4 cm. The 242Pu5+ has been transported down the beam line with a deflection voltage equal to 0 V. Fig. 6 shows the equivalent plot in the bending (X) plane for 239Pu5+ at the same terminal voltage, with appropriate voltages applied to the deflector plates. Table 1 summarises the voltages required, as determined from the COSY Infinity simulations. Implementation of these steerers has been deferred pending full commissioning of the rest of the system. In
the meantime, actinide measurements will continue to be performed by changing the terminal voltage and Wien filter voltage, although the injection of the different isotopes into the accelerator will be controlled by changing the pre-acceleration voltage rather than the injection magnet. 3.4. Sequencing The sequencing system, beam current integrators, and data acquisition system for the detectors has been purchased from NEC. The data acquisition system incorporates Pixie digital pulse processing modules (manufactured by Xia LLC) capable of handling up to 8 parameters. Given the 30 ms settling time of the pre-acceleration voltage, cycling between isotopes will be at 2 Hz, which should be faster than most ion source fluctuations
480
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Fig. 5. Simulation with COSY Infinity of the 242Pu5+ beam at 23.846 MeV. The beam starts at the object point of the analyzing magnet with a ±1.5 mm width and an angular divergence of 3 mrad. The distance from the object slits of the analysing magnet to the detector is 18 m, and the gap in the analysing magnet in the bending (X) plane is 5.9 cm. ES is an electrostatic deflector, AM is the analysing magnet, SM the switching magnet, MQ is a magnetic quadrupole doublet, WF is the Wien filter, and D is the detector.
while still allowing 90% of the time to be spent in measuring the rare isotope. Initially, the Pixie modules will be operated essentially as peak-sensing ADCs, but as experience is gained, we will
explore their potential in providing additional discrimination against background and pile-up events by using them in pulse-shape sampling mode.
481
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Fig. 6. Simulation with COSY Infinity of the 239Pu5+ transport beam at 23.846 MeV and with the appropriate voltages on the deflector plates and all the other elements set for the 242Pu5+ reference beam (see text). Conditions and notation are the same as for Fig. 5. Only the X (bending) plane is shown. The Y plane is identical to that in Fig. 5. The black line (BL) is the trajectory of the 242Pu5+.
Table 1 Optimized values for the voltages on the deflector plates before and after the analyzing magnet (AM) and switching magnet (SM) for 240Pu5+ and 239Pu5+ at a terminal voltage of 4 MV [2]. Voltages are applied symmetrically to both plates. X and A are the distance and angle respectively of the central ray from the axis, at the image point of the analyzing magnet and at the detector position for the analyzing magnet and switching magnet respectively. Note that the deflectors before and after the analyzing magnet do not change either the position or angle of the beam at the image point of the magnet, but there is a small change in angle after the switching magnet. 240
239
Pu
Analyzing magnet (kV) Before
X (mm)
A (mrad)
After
±8.000 ±8.000 Switching magnet (kV) Before
After
2.400
±2.400
Pu
Analyzing magnet (kV) Before
0 X (mm)
0
0 A (mrad)
0.2
4. Summary and conclusions The present status of the upgrade of the 14UD accelerator at ANU to enable fast cycling is as follows.
Low-energy side Beams of the different isotopes are switched into the accelerator by changing the energy of the beams from the ion source. This is achieved with a TREK ±10 kV fast amplifier in series with a Glassman 200 kV supply. Settling times after a voltage change of the TREK are 30 ms. A beam chopper immediately after the injection magnet acts as a blanking steerer to prevent the beam entering the accelerator during this interval. Because the beam energies of the different isotopes in the electrostatic quadrupole triplet lens before the accelerator are different, it is also necessary
X (mm)
A (mrad)
0 X (mm)
0 A (mrad)
After
±11.950 ±11.950 Switching magnet (kV) Before
After
3.650
±3.650
0
0.4
to vary the voltages on this lens from isotope to isotope. This is achieved with a set of six TREK high-voltage amplifiers. High energy side A new vacuum box has been installed in the high-energy analyzing magnet that will accommodate mass differences up to 10% between the rare and stable isotopes. Off-axis Faraday cups have been installed at a distance of 80 cm from the analyzing magnet. COSY Infinity simulations of beam displacements and widths have been performed to define the optimal sizes of the cups, and verified with an actual beam. In addition, COSY Infinity simulations have been performed to determine the trajectories of different plutonium isotopes when fast-switched steerers before and after the analyzing and switching magnets are employed to direct the different isotopes to the detector. Implementation of this capability has, however, been deferred until the rest of the system has been fully commissioned.
482
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Acknowledgements Funding for these developments was provided by the Australian Government through the Heavy Ion Accelerators Education Investment Fund (EIF) Project. MDC is indebted to his father, Prof. Nicola De Cesare of the II University of Naples, for invaluable discussions, as well as Prof. Lucio Gialanella of the II University of Naples and Dr. Detlef Rogalla of the Ruhr-Universität Bochum for suggestions. References [1] L.K. Fifield, S.G. Tims, T. Fujioka, W. Hoo, S. Everett, Accelerator mass spectrometry with the 14UD accelerator at the Australian National University, Nucl. Instr. Meth. Phys. Res. B 268 (2010) 858.
[2] L.K. Fifield, Accelerator mass spectrometry of the actinides, Quat. Geochronol. 3 (2008) 276. [3] D.C. Weisser, L.K. Fifield, M. De Cesare, S.G. Tims, N.R. Lobanov, G.G. Crook, D. Tsifakis, T.B. Tunningley, Injection optics for fast mass switching for accelerator mass spectrometry, AIP Conf. Proc. 1515 (2013) 464. [4] K. Makino, M. Berz, COSY INFINITY version 8, Nucl. Instr. Meth. Phys. Res. A 427 (1999) 338. [5] B. Zorko, D.P. Child, M.A.C. Hotchkis, A fast switching electrostatic deflector system for actinide isotopic ratio measurements, Nucl. Instr. Meth. Phys. Res. B 268 (2010) 827. [6] S.H. Sie, D.A. Sims, T.R. Niklaus, G.F. Suter, A fast bouncing system for the highenergy end of AMS, Nucl. Instr. Meth. Phys. Res. B 172 (2000) 268. [7] M. De Cesare, D.C. Weisser, L.K. Fifield, T.B. Tunningley, N.R. Lobanov, A novel beam focus control at the entrance to the ANU 14UD accelerator, in: EPJ Web Conf. 63 (2013) 1–4. id: 03008.
Contents lists available at ScienceDirect
Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb
A new fast-cycling system for AMS at ANU M. De Cesare ⇑, L.K. Fifield, D.C. Weisser, D. Tsifakis, A. Cooper, N.R. Lobanov, T.B. Tunningley, S.G. Tims, A. Wallner Department of Nuclear Physics, Research School of Physics and Engineering, Australian National University, ACT 2601, Canberra, Australia
a r t i c l e
i n f o
Article history: Received 15 December 2014 Received in revised form 20 May 2015 Accepted 20 May 2015 Available online 29 May 2015 Keywords: Accelerator mass spectrometry Fast-cycling system Actinides
a b s t r a c t In order to perform higher precision measurements, an upgrade of the ANU accelerator is underway. Fast switching times on the low-energy side, with maximum settling times of 30 ms, are achieved by holding the injector magnet field constant while changing the energy of the different isotopes by changing the pre-acceleration voltage after the ion source. Because ions of the different isotopes then have different energies before injection, it is necessary to adjust the strength and steering of the electrostatic quadrupole lens that focusses the beam before entry into the accelerator. First tests of the low-energy system will be reported. At the high energy end, a larger vacuum box in the analyzing magnet has been designed, manufactured and installed to allow the transport of differences in mass as large as 10% at constant terminal voltage. For the cases where more than one isotope must be transported to the detector an additional refinement is necessary. If the accelerator voltage is to be kept constant, then the trajectories of the different isotopes around both the analyzing and switching magnets must be modified. This will be achieved using bounced electrostatic steerers before and after the magnets. Simulations have been performed with the ion optic code COSY Infinity to determine the optimal positions and sizes of these steerers. Ó 2015 Published by Elsevier B.V.
1. Introduction Due to the exigencies of sharing a large accelerator with an active nuclear physics program, and limitations imposed by existing hardware, AMS at the Australian National University [1] (ANU) has for many years operated in a sequential slow-cycling mode. Counting times for the rare isotope are typically two to five minutes, and are dictated by the requirement to maximize the fraction of time spent on counting the rare isotope(s) when switching times between isotopes are 10–15 s. Ion source output can, however, vary appreciably during these counting times, which limits the precision that can presently be achieved to 3%. In order to perform higher precision measurements, an upgrade of the ANU accelerator is underway to allow a fast-cycling procedure. Here we describe the various components of the upgrade and its present status. 2. Upgrade of the ANU accelerator and actinide measurements The ANU AMS system is based on a 15 MV tandem accelerator. As high energy is required to apply certain techniques of isobar ⇑ Corresponding author at: Italian Aerospace Research Centre – CIRA, Via Maiorise, 81043 Capua (CE), Italy. E-mail address: [email protected] (M. De Cesare). http://dx.doi.org/10.1016/j.nimb.2015.05.028 0168-583X/Ó 2015 Published by Elsevier B.V.
separation effectively, this makes the ANU tandem particularly well-suited for the heavier isotopes with severe isobaric interferences e.g., 36Cl and 53Mn, but it is sufficiently flexible that it can also be run at lower voltages 4 MV for actinide measurements [1,2]. A schematic layout of the ANU 15 MV tandem facility is shown in Fig. 1. The caesium sputter ion source is a 32-sample MC-SNICS. Negative ions from this source are pre-accelerated and mass-analysed by the mass-energy product 56 MeV amu/e2 injection magnet, which allows high resolution mass analysis for all stable isotopes in the periodic table. The injection beam line also features a fast (rise-time 50 ns) electrostatic chopper that allows attenuation of beam currents that would be too high for injection into the tandem accelerator (e.g. 35Cl) or counting rates that would be too high for the detector (e.g. 234U). An electrostatic quadrupole triplet with steering capability focusses the beams to a waist inside the accelerator tank. This waist is the object point for the lens formed by the electric field at the entrance to the acceleration tube which then focusses the beams into the stripper canal in the high voltage terminal. The 14UD accelerator is capable of voltages up to 15 MV and is routinely operated above 14 MV. Both a gas stripper and a foil stripper are available at the terminal. The double focusing high-energy analyzing magnet has a radius of 1.27 m and a mass-energy product of 225 MeV amu/e2 at its maximum field of 1.7 T. Its vacuum box has an opening in the bending direction of
476
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Fig. 1. Schematic layout of the ANU 14UD accelerator and the 15° beam line that is used inter alia for AMS of actinides showing actual and planned upgrades for fast-cycling AMS operation. In the actinides beam line A denotes a selectable 3 or 6 mm diameter aperture. The accelerator and analysing magnet are vertical, while the switching magnet bends in the horizontal plane. The upgraded part includes the ±10 kV TREK power supply in the pre-acceleration to switch between different isotopes, the new analyzing magnet chamber and the offset Faraday cups after the analyzing magnet. The deflector plates before and after the analyzing and switching magnets will be installed at a later date. Note that the ionization chamber is removed when either the time-of-flight system or the gas-filled magnet are being used.
only 59 mm at the exit from the magnet. This is insufficient to allow 35Cl, for example, to be transmitted at the same time as 36 Cl when the latter is on the central axis, and has been a major inhibiting factor to the development of a fast-cycling capability. Two beam lines are available for AMS, one equipped with a multi-element gas ionization detector, the other with a 6 m time-of-flight system, a gas-filled magnet, and an ionization
detector. Wien filters on both beam lines provide the final analysis stage to remove ions with the same mE/q2 but different E/q. In the slow-cycling mode, different isotopes are injected into the accelerator by changing the magnetic field in the injection magnet. On the high-energy side, the analyzing and switching magnets are held fixed while the terminal voltage as well as the electric field in the Wien filter are changed in order to transmit
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
all the isotopes to the detector station. For example, 35Cl, 37Cl and 36 Cl are injected in a cycle in which the beam currents of 35Cl and 37 Cl are measured in a Faraday cup immediately in front of the gas ionization detector for 10 s each, while the 36Cl ions are counted in the detector for 5 min. In the case of plutonium, 242Pu, 240Pu and 239 Pu are all counted in the detector for typically 1, 3 and 2 min respectively. For the fast-cycling system presently being implemented, fast switching times on the low-energy side will be achieved by holding the injector magnet field constant while changing the energy of the different isotopes by changing the pre-acceleration voltage after the ion source. This normally operates at 150 kV. The incremental voltages required for fast switching between isotopes will be supplied by a fast high-voltage amplifier (±10 kV from TREK) in series with the Glassman 200 kV supply [3]. Because the energy of the injected beam is changed, it is also necessary to adjust the strength of an electrostatic quadrupole triplet before injection, and this will be achieved with six additional TREK high-voltage amplifiers. At the high energy end a larger vacuum box in the analyzing magnet has been designed, manufactured and installed to allow the transport of isotopes differing in mass by up to 10% (e.g. 9Be and 10Be). Currents of the stable beams (e.g. 35Cl and 37 Cl) will be measured in offset Faraday cups after the analyzing magnet, and the appropriate vacuum housing and Faraday cups have been designed and constructed on the basis of detailed beam optics calculations using the code COSY Infinity [4]. These cups are attached to motorized linear drives so that they can be positioned appropriately for the wide range of isotopes measured with the 14UD accelerator. For the cases where more than one isotope must be transported to the detector (e.g. 239,240,242Pu or 233,236U) an additional refinement is necessary. If the accelerator voltage is to be kept constant, then the trajectories of the different isotopes around both the analyzing and switching magnets must be modified. This will be achieved using bounced electrostatic steerers before and after the magnets, as implemented for example at ANSTO for actinide measurements and formerly at the Sydney CSIRO facility for ‘super-SIMS’ applications [5,6]. Simulations have been performed with COSY Infinity to determine the optimal positions and sizes of these steerers.
477
160 kV to 140 kV. Clearly, the response time is long compared to the 100 ls rise time of the TREK amplifier due to the time taken for the Glassman supply to adjust to the changed current demand. The manufacturer quotes a time constant of 50 ms for the response of the Glassman supply to a demand for a voltage change, so it is not surprising that the response time of the voltage stabilisation circuit to a change in current is of the same order. Of course, this 20 kV swing is the extreme case, and actual changes will be substantially smaller. Switching between 35Cl and 37Cl, for example, will require a change from about 4 kV to +4 kV where the response is faster. Nevertheless, it is necessary to wait for 30 ms for the accelerating voltage to stabilise before performing the current measurements of the stable beams, or starting the counting of the rare isotope. During this 30 ms, the low-energy chopper will be employed as a ‘blanking steerer’ to sweep the beam onto the image slits of the injection magnet so that it is not injected into the accelerator. In order to explore whether the response time could be shortened, an additional current drain from the Glassman supply was effected by adding a 200 MO resistor chain from high voltage to ground. As can be seen in Fig. 2, this steepens the rising edge of the pulse, but it subsequently takes longer to return to the baseline so was not pursued further. 3.2. Low-energy focus control The quadrupole triplet lens that forms the final waist before the beam enters the low-energy accelerator tube is electrostatic. Hence, its strength for given applied voltages depends on ion energy. Since the various ions have different energies in this lens, it is necessary to adjust the applied voltages from one ion species to another. In [3] preliminary measurements with a DC beam of 28 Si at a terminal voltage of 12.7 MV confirmed that the transmission decreased when the injection voltage was reduced or increased by 4 kV from the nominal 150 kV. In both cases, full transmission could be restored by small changes in the strength of the lens only. Since the lens incorporates steering as well as focussing, however, it requires six high-voltage power supplies and adjustments to all six are required in order to change the strength of the lens. The response times of the six Glassman power supplies that previously supplied these voltages were too slow to be compatible with the fast-switching sequence, and have been
3. Simulations and experimental results 3.1. Low-energy side of the accelerator Fast switching between isotopes in modern AMS facilities is almost universally accomplished by modifying the beam energy though the 90° injection magnet by applying voltages of several kV to the insulated vacuum box of the magnet. This option was not practicable at the ANU, both because of the presence of the fast beam chopper immediately after the injection magnet, and because the magnet must be rotated between two ion sources. So instead, the novel alternative of pulsing the voltage on the high voltage ion source deck is being implemented using a TREK 1040 A, ±10 kV power supply connected in series with the Glassman 200 kV power supply that presently supplies the pre-acceleration voltage of typically 150 kV. The TREK power supply is controlled via an analog ±10 V signal which is transferred from ground potential via a fiber optic cable [3]. It will be controlled by a sequencer made by NEC. A series of tests were performed in order to quantify the effectiveness of the TREK power supply in rapidly changing the ion source deck voltage. The results are summarised in Fig. 2, which shows the response of the ion source deck when the TREK amplifier was switched from +10 kV to 10 kV in order to change the accelerating voltage from
Fig. 2. Oscillograms showing the response of the voltage of the ion source deck when the Glassman high voltage supply is set to 150 kV and the TREK supply is switched at time = 0 ms from +10 to 10 kV (the chassis of the TREK is connected to the deck), and from 10 to +10 kV at time = 50 ms. The trace with the faster rise time was obtained with an additional 200 MO load on the Glassman supply (see text). The width of the ‘measuring time’ has been exaggerated for clarity.
478
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
replaced by six TREK 10 kV fast amplifiers. These are unipolar, with four being negative and two positive. The focussing at the entrance to the low-energy acceleration tube of the 14UD accelerator is controlled by a novel tube-entrance lens [7] that uses a 150 kV power supply. Its strength for a given voltage also depends on the energy of the injected ions, and in principle it should also be adjusted from isotope to isotope. It was not practicable to change this voltage, but the changes in focussing may be compensated by moving the beam-waist closer to or further from the tube-entrance lens with the quadrupole triplet. The strengths of the quadrupole triplet therefore differ slightly from what would be required to focus all isotopes to the same waist position, and are set empirically to optimise transmission for each isotope. 3.3. High-energy side of the accelerator The new analyzing magnet chamber, Fig. 3, and the subsequent box to house the off-axis Faraday cups were designed on the basis of simulations with the code COSY Infinity. In particular, the chamber and Faraday cup box are designed to accommodate off-axis 9Be ions when 10Be of the same energy is on the central trajectory. Accordingly the exit aperture is 180.6 mm in the bending plane, which is a factor of three larger than the 59.0 mm of the original vacuum chamber. The inside gap of the vacuum chamber parallel to the magnetic field is 23.75 mm, and the aperture in the bending plane at the entrance is 59.0 mm. In order to define the dimensions of the offset Faraday cups, we have simulated the displacement between the commonly-used isotopes that are the most difficult to separate, i.e. between the abundant 238U and the rare 236U and between the abundant 56Fe, 57 Fe and the rare 60Fe. The distance between the exit from the magnet and the Faraday cups is constrained to 80 cm by the need to rotate the analyzing magnet between two target rooms, since a beam line valve is required to isolate the magnet vacuum from
the downstream system when the magnet is being rotated. The off-axis Faraday cups are therefore not at the image plane of the 127 cm radius magnet and it is important to know the beam dimensions at their actual positions. The results of the simulation obtained with the COSY Infinity beam optics program are shown in Fig. 4. At the object point of the analyzing magnet the beam is assumed to have a full width of ±1.5 mm and a maximum angular divergence of 3 mrad which is the worst case scenario obtained with uranium. It is clear that 238U5+ and 236U5+ ions are not sufficiently separated at 80 cm from the magnet, but would be well separated if an additional off-axis cup were placed at 2 m from the magnet. This would be possible, since even at 2 m from the magnet the 238U ions are only 2 cm off axis and hence could be transported through the beam line valves and other apertures. In the case of the Fe isotopes, 56Fe and 57Fe are seen to be well separated from each other and from 60Fe even at 80 cm from the magnet. The beam full widths are 12 mm and the separation between 56 Fe and 57Fe is 24 mm, and hence the individual isotopes are comfortably accommodated in Faraday cups with inner and outer diameters of about 20 and 22 mm respectively. Tests with a 35Cl beam indicate that the beam size is indeed comfortably within the cup. Note that the 12 mm full width of the 60Fe beam is that at 80 cm from the magnet, and assumes an angular spread of 3 mrad, which is probably a significant overestimate. At the image point of the magnet, 254 cm from the magnet exit, the beam full width is 3 mm. Plutonium isotopes, extracted as molecular ions (xPu16O ) [2], must all be counted in the ionization chamber. In order to transmit 239,240,242 Pu sequentially to the detector at a fixed terminal voltage, the trajectories around the analyzing and switching magnets must be modified with electrostatic steerers, the voltages of which can be switched from beam to beam. Provision has been made to house such a deflector after the analyzing magnet by incorporating a deflector chamber immediately after the magnet as shown in Fig. 3. Similar chambers are envisaged before the analyzing magnet, and before and after the switching magnet. A simulation of
Fig. 3. The new magnet chamber, the deflector chamber and the chamber that houses three offset movable Faraday cups for the abundant isotopes, see text. The blue central ray represents the rare selected isotope, the two red (upper) rays represent lighter isotopes, while the green (lower) ray represents a heavier isotope. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
479
Fig. 4. Beam positions and widths after the analyzing magnet at the indicated distances from the magnet exit. The upper plot shows uranium isotopes, while the lower plot shows iron isotopes. It is assumed that, at the object point of the analysing magnet, the beam is uniformly distributed over a full width of ±1.5 mm and an angular spread of 3 mrad.
the beam trajectories for xPu isotopes from the object point of the analyzing magnet to the Ionization Chamber (Fig. 1) in the actinides beam line has been performed with the COSY Infinity beam optics program. Fig. 5 shows the result for a 23.846 MeV 242Pu5+ beam (4 MV terminal voltage with an injection energy of 100 keV), in the bending plane, X, and vertical plane, Y. The deflector plates are positioned 42 cm before and after the analysing and switching magnet, and have a length of 12 cm and a gap of 4 cm. The 242Pu5+ has been transported down the beam line with a deflection voltage equal to 0 V. Fig. 6 shows the equivalent plot in the bending (X) plane for 239Pu5+ at the same terminal voltage, with appropriate voltages applied to the deflector plates. Table 1 summarises the voltages required, as determined from the COSY Infinity simulations. Implementation of these steerers has been deferred pending full commissioning of the rest of the system. In
the meantime, actinide measurements will continue to be performed by changing the terminal voltage and Wien filter voltage, although the injection of the different isotopes into the accelerator will be controlled by changing the pre-acceleration voltage rather than the injection magnet. 3.4. Sequencing The sequencing system, beam current integrators, and data acquisition system for the detectors has been purchased from NEC. The data acquisition system incorporates Pixie digital pulse processing modules (manufactured by Xia LLC) capable of handling up to 8 parameters. Given the 30 ms settling time of the pre-acceleration voltage, cycling between isotopes will be at 2 Hz, which should be faster than most ion source fluctuations
480
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Fig. 5. Simulation with COSY Infinity of the 242Pu5+ beam at 23.846 MeV. The beam starts at the object point of the analyzing magnet with a ±1.5 mm width and an angular divergence of 3 mrad. The distance from the object slits of the analysing magnet to the detector is 18 m, and the gap in the analysing magnet in the bending (X) plane is 5.9 cm. ES is an electrostatic deflector, AM is the analysing magnet, SM the switching magnet, MQ is a magnetic quadrupole doublet, WF is the Wien filter, and D is the detector.
while still allowing 90% of the time to be spent in measuring the rare isotope. Initially, the Pixie modules will be operated essentially as peak-sensing ADCs, but as experience is gained, we will
explore their potential in providing additional discrimination against background and pile-up events by using them in pulse-shape sampling mode.
481
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Fig. 6. Simulation with COSY Infinity of the 239Pu5+ transport beam at 23.846 MeV and with the appropriate voltages on the deflector plates and all the other elements set for the 242Pu5+ reference beam (see text). Conditions and notation are the same as for Fig. 5. Only the X (bending) plane is shown. The Y plane is identical to that in Fig. 5. The black line (BL) is the trajectory of the 242Pu5+.
Table 1 Optimized values for the voltages on the deflector plates before and after the analyzing magnet (AM) and switching magnet (SM) for 240Pu5+ and 239Pu5+ at a terminal voltage of 4 MV [2]. Voltages are applied symmetrically to both plates. X and A are the distance and angle respectively of the central ray from the axis, at the image point of the analyzing magnet and at the detector position for the analyzing magnet and switching magnet respectively. Note that the deflectors before and after the analyzing magnet do not change either the position or angle of the beam at the image point of the magnet, but there is a small change in angle after the switching magnet. 240
239
Pu
Analyzing magnet (kV) Before
X (mm)
A (mrad)
After
±8.000 ±8.000 Switching magnet (kV) Before
After
2.400
±2.400
Pu
Analyzing magnet (kV) Before
0 X (mm)
0
0 A (mrad)
0.2
4. Summary and conclusions The present status of the upgrade of the 14UD accelerator at ANU to enable fast cycling is as follows.
Low-energy side Beams of the different isotopes are switched into the accelerator by changing the energy of the beams from the ion source. This is achieved with a TREK ±10 kV fast amplifier in series with a Glassman 200 kV supply. Settling times after a voltage change of the TREK are 30 ms. A beam chopper immediately after the injection magnet acts as a blanking steerer to prevent the beam entering the accelerator during this interval. Because the beam energies of the different isotopes in the electrostatic quadrupole triplet lens before the accelerator are different, it is also necessary
X (mm)
A (mrad)
0 X (mm)
0 A (mrad)
After
±11.950 ±11.950 Switching magnet (kV) Before
After
3.650
±3.650
0
0.4
to vary the voltages on this lens from isotope to isotope. This is achieved with a set of six TREK high-voltage amplifiers. High energy side A new vacuum box has been installed in the high-energy analyzing magnet that will accommodate mass differences up to 10% between the rare and stable isotopes. Off-axis Faraday cups have been installed at a distance of 80 cm from the analyzing magnet. COSY Infinity simulations of beam displacements and widths have been performed to define the optimal sizes of the cups, and verified with an actual beam. In addition, COSY Infinity simulations have been performed to determine the trajectories of different plutonium isotopes when fast-switched steerers before and after the analyzing and switching magnets are employed to direct the different isotopes to the detector. Implementation of this capability has, however, been deferred until the rest of the system has been fully commissioned.
482
M. De Cesare et al. / Nuclear Instruments and Methods in Physics Research B 361 (2015) 475–482
Acknowledgements Funding for these developments was provided by the Australian Government through the Heavy Ion Accelerators Education Investment Fund (EIF) Project. MDC is indebted to his father, Prof. Nicola De Cesare of the II University of Naples, for invaluable discussions, as well as Prof. Lucio Gialanella of the II University of Naples and Dr. Detlef Rogalla of the Ruhr-Universität Bochum for suggestions. References [1] L.K. Fifield, S.G. Tims, T. Fujioka, W. Hoo, S. Everett, Accelerator mass spectrometry with the 14UD accelerator at the Australian National University, Nucl. Instr. Meth. Phys. Res. B 268 (2010) 858.
[2] L.K. Fifield, Accelerator mass spectrometry of the actinides, Quat. Geochronol. 3 (2008) 276. [3] D.C. Weisser, L.K. Fifield, M. De Cesare, S.G. Tims, N.R. Lobanov, G.G. Crook, D. Tsifakis, T.B. Tunningley, Injection optics for fast mass switching for accelerator mass spectrometry, AIP Conf. Proc. 1515 (2013) 464. [4] K. Makino, M. Berz, COSY INFINITY version 8, Nucl. Instr. Meth. Phys. Res. A 427 (1999) 338. [5] B. Zorko, D.P. Child, M.A.C. Hotchkis, A fast switching electrostatic deflector system for actinide isotopic ratio measurements, Nucl. Instr. Meth. Phys. Res. B 268 (2010) 827. [6] S.H. Sie, D.A. Sims, T.R. Niklaus, G.F. Suter, A fast bouncing system for the highenergy end of AMS, Nucl. Instr. Meth. Phys. Res. B 172 (2000) 268. [7] M. De Cesare, D.C. Weisser, L.K. Fifield, T.B. Tunningley, N.R. Lobanov, A novel beam focus control at the entrance to the ANU 14UD accelerator, in: EPJ Web Conf. 63 (2013) 1–4. id: 03008.