Simulation and optimization of a 10 A electron gun ...

Simulation and optimization of a 10 A electron gun with electrostatic compression for the electron beam ion source A. Pikin, E. N. Beebe, and D. Raparia Citation: Rev. Sci. Instrum. 84, 033303 (2013); doi: 10.1063/1.4793773 View online: http://dx.doi.org/10.1063/1.4793773 View Table of Contents: http://rsi.aip.org/resource/1/RSINAK/v84/i3 Published by the American Institute of Physics.

Additional information on Rev. Sci. Instrum. Journal Homepage: http://rsi.aip.org Journal Information: http://rsi.aip.org/about/about_the_journal Top downloads: http://rsi.aip.org/features/most_downloaded Information for Authors: http://rsi.aip.org/authors

Downloaded 09 Apr 2013 to 130.199.3.165. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://rsi.aip.org/about/rights_and_permissions

REVIEW OF SCIENTIFIC INSTRUMENTS 84, 033303 (2013)

Simulation and optimization of a 10 A electron gun with electrostatic compression for the electron beam ion source A. Pikin,a) E. N. Beebe, and D. Raparia Brookhaven National Laboratory, Upton, New York 11973, USA

(Received 19 December 2012; accepted 15 February 2013; published online 11 March 2013) Increasing the current density of the electron beam in the ion trap of the Electron Beam Ion Source (EBIS) in BNL’s Relativistic Heavy Ion Collider facility would confer several essential benefits. They include increasing the ions’ charge states, and therefore, the ions’ energy out of the Booster for NASA applications, reducing the influx of residual ions in the ion trap, lowering the average power load on the electron collector, and possibly also reducing the emittance of the extracted ion beam. Here, we discuss our findings from a computer simulation of an electron gun with electrostatic compression for electron current up to 10 A that can deliver a high-current-density electron beam for EBIS. The magnetic field in the cathode-anode gap is formed with a magnetic shield surrounding the gun electrodes and the residual magnetic field on the cathode is (5 ÷ 6) Gs. It was demonstrated that for optimized gun geometry within the electron beam current range of (0.5 ÷ 10) A the amplitude of radial beam oscillations can be maintained close to 4% of the beam radius by adjusting the injection magnetic field generated by a separate magnetic coil. Simulating the performance of the gun by varying geometrical parameters indicated that the original gun model is close to optimum and the requirements to the precision of positioning the gun elements can be easily met with conventional technology. © 2013 American Institute of Physics. [http://dx.doi.org/10.1063/1.4793773] I. INTRODUCTION

BNL’s Electron Beam’s Source (EBIS) operates in the Relativistic Heavy Ion Collider (RHIC) accelerating facility supplying a broad variety of highly charged ions for highenergy acceleration in RHIC- and for NASA applications, requiring ions with Booster energy determined by its maximum rigidity of 16 Tm.1, 2 RHIC’s requirements for the charge state of accelerated ions have been satisfied with an EBIS electron beam’s current density of 500 A/cm2 in the ion trap.3, 4 Such current density is achieved via a combination of a highperformance thermionic cathode and the subsequent compression of the electron beam by the magnetic field of the main superconducting solenoid.5–8 The required charge state of the Au ions from the EBIS is 32+, and normally, it takes 40–60 ms to reach this charge state therein. Presently, for our NASA experiments, there is a demand to increase the maximum ion energy after the Booster. One way to do so is to increase the charge state of ions generated in the EBIS. However, such an increase necessitates a larger value for the ionization factor jτ , viz., the product of the density of the electron beam’s current, j, and the confinement time, τ . The extended retention of the ions in a trap increases the number of accumulated residual-gas ions that can displace the injected working ions. A further rise of the electron beam’s current density in the ion trap using the existing concept of the immersed electron gun by increasing the cathode emission current density, or by reducing the magnetic field on the cathode hardly is practical because of the steep shortening of the cathode’s life time at higher temperatures and the increase of radial oscillations of the electron beam, entailing degradation of the beam’s quality at a lower magnetic field on the cathode. a) [email protected].

0034-6748/2013/84(3)/033303/5/$30.00

Instead, adopting the concept of a combined electrostaticand magnetostatic-compression looks more promising. This idea was used with various degrees of success in early EBIS devices,9, 10 and was quite successfully used in Electron Beam Ion Traps (EBITs)11, 12 with moderate electron currents. It can provide an electron beam current with density 10–20 higher than that currently used in our BNL EBISes. Generally, EBIS operation requires the electron beam’s current to be controllable over a wide range in different regimes. The existing immersed electron gun allows such control without degrading the beam’s quality. One requirement of a new gun is that it should deliver a high-current-density laminar electron beam with a current range of 1–10 A. Since neither high-currentdensity EBIS nor EBIT devices have not yet been tested with multiampere electron currents, our ability to run the EBIS with high-compression electron gun in a broad range of electron current support our efforts to find a limit for a stable trap operation, provided that it exists within this range of current. The theory of Brillouin flow denotes a strong dependence of the final radius of an electron beam in a magnetic field on the residual magnetic field on the cathode. According to this theory, the radius of the electron beam containing 80% of the total current, re , in the magnetic field, B, can be calculated as follows:13 re = rB

BC4 rC4 8kT rC2 + 0.5 + 0.5 1 + 4 mη2 rB B 2 B 2 rB4

0.5 0.5 ,

(1) wherein kT is the electron energy on the cathode determined by the cathode temperature T, rc the cathode radius, Bc the magnetic field on the cathode, m the electron mass, and η ratio of the electron charge to mass. The Brillouin radius, rB , can

84, 033303-1

© 2013 American Institute of Physics


033303-2

Pikin, Beebe, and Raparia

Rev. Sci. Instrum. 84, 033303 (2013)

FIG. 2. Simulation model with magnetic-field’s distribution optimized for a 10 A electron beam current.

FIG. 1. Schematic of the electron gun with electrostatic focusing.

be calculated using a convenient formula,14 √ 0.015 Iel rB = , BEel0.25

(2)

where Iel is the electron current (A), B is the magnetic field (T), Eel the electron energy (keV), and resulting rB is in cm. Formula (1) assumes that there is no radial scalloping of the beam. We paid special attention to minimizing the amplitude of radial oscillations of the generated electron beam because they will determine the final diameter of the beam and the maximum perveance of the electron beam in the ion trap. II. SIMULATION MODEL

The cathode of the existing electron gun of RHIC EBIS is located in the magnetic field produced by the combination of an independently controlled external “warm” magnet coil, and the main superconducting solenoid5 with the contribution from the coil dominating. A new electron gun will occupy the same space as the existing gun inside a slightly modified vacuum chamber. Figure 1 is a general schematic of the gun elements with their main dimensions. The gun assembly with magnetic shield is located inside the external gun coil, such that the right edge of the front plate of this shield is positioned with its aperture in the median plane of the gun coil. The magnetic field within the proposed Pierce-type electron gun is formed by a combination of the external magnetic field and the magnetic shield surrounding the cathodeanode gap. The magnetic shield is electrically connected to the anode. In this geometry, the distribution of the magnetic field in the cathode-anode gap primarily is determined by the

magnetic shield’s front plate with the aperture and by the gun coil’s current. This shield is long enough to effectively suppress the contribution to the magnetic field on the cathode from the shield’s opening on the side opposite the front plate. Figure 2 depicts the simulated distribution of the magnetic field in the gun region. For our computer simulations of the electron gun, we used a package of programs from FieldPrecision Co.15 The magnetic shield reduces the magnetic field on the cathode to a few Gs. No additional coil for cancelling the residual magnetic field on the cathode is planned presently because of the cost, complexity, and reliability considerations. Besides, we do not consider that the residual magnetic field, which limits the maximum attainable current density in the center of the trap, is a disadvantage for our EBIS application. In fact, having a small residual magnetic field on the cathode, and therefore, a weaker dependence of the radius of the electron beam on the neutralization factor of the electron beam in the trap can be considered a plus, because it stabilizes the final ion charge state spectrum and the final potential on the trap’s axis. The magnetic-field’s distribution in the Test EBIS has a “valley” between the gun coil and the main superconducting solenoid, with minimum magnetic field of a 350 gauss. With such distribution of the magnetic field, the electron beam, upon exit from gun, expands in this “valley” before becoming compressed in the main solenoid. Even with larger expansion factor for a Brillouin electron beam than for an immersed flow beam, the maximum diameter of the former in the minimum of the magnetic field is still almost 2 times smaller than of our existing electron beam. Therefore, the existing structure of the drift tubes in this transition area is expected to accommodate this expanded electron beam without modification. III. RESULTS OF SIMULATIONS A. Optimization of the gun’s geometry

A choice of the specific electrostatic geometry for the electron gun was determined by the operating range of the


033303-3


FIG. 3. Simulated electron trajectories for optimized gun parameters. Iel = 10.0 A, Ua = 49.2 kV, Binj = 2.1 kGs.

existing gun coil that can generate maximum magnetic field Bcoil_max = 2.5 kGs. The electron trajectories were simulated on a distance of up to 160 mm from the cathode in a falling magnetic field of the EBIS Test stand, with the radial size of the electrostatic mesh at 0.070 mm. Figure 3 illustrates the simulated electron trajectories for a nominal electron current Iel = 10.0 A and the optimized gun parameters. The simulated perveance of the electron gun is pgun = 0.92 × 10−6 A/V1.5 . Our simulations did not demonstrate advantage of varying the Wehnelt electrode’s voltage with respect to the cathode. We have not investigated the effect of the cathode edge; a beam halo originating from this edge is expected. In our gun design, the Wehnelt electrode remains isolated from the cathode for fine tuning of the beam so to minimize a possible halo. The radial gap between the Wehnelt electrode and the cathode is 0.2 mm. Initially, the front plate of the magnetic shield with the aperture was positioned with its entrance plane on a crossover of the electron beam simulated without a magnetic field, and the initial radius of the diaphragm was taken as 4.7 mm. Optimizing the gun’s geometry included changing of the position of the shield’s front plate with respect to a fixed cathode/anode gap, of the radius of the magnetic shield’s aperture, and of the cathode’s position with respect to the anode/magnetic shield

FIG. 4. Dependence of the minimum radial oscillations (dr_min) and of the residual magnetic field on the cathode (B_cath) on the radius of the aperture in the front plate of the magnetic shield for 10 A electron beam.

Rev. Sci. Instrum. 84, 033303 (2013)

FIG. 5. Dependences of minimum radial beam oscillations on the shift of the cathode with respect to the anode (dr_min_cath), and on the shift of vertical plate of the magnetic shim (dr_min_shim) from their initial positions.

assembly. A value of minimum amplitude of the radial oscillations, dr_min for each case, was determined from several runs with different injection magnetic fields. Figures 4 and 5 present the results of this optimization. These graphs illustrate that, for the most part, the original geometry of the gun was close to optimum, except of the position of the magnet shield’s front plate that should be shifted in a positive direction (to the right, according to Fig. 1) by 0.5 mm. If the initial amplitude of radial oscillations of 0.1 mm is acceptable, the tolerance of the axial cathode positioning lies in a range of −0.6 mm to +0.4 mm, and the tolerance of the magnet shield front plate positioning is ±0.7 mm. B. “Broadband” operation of the gun

A controllable external magnetic field allows us to use this gun in a broad range of electron current. By tuning the injection magnetic field, we can vary the electron beam’s current and maintain the matching conditions of the electric- and

FIG. 6. Dependence of amplitude of the electron beam’s radial oscillations (dr) on the injection magnetic field, B, for electron beam currents of 1.0 A and 10.0 A.


033303-4


FIG. 7. Simulated dependences of the optimum injection magnetic field (B_opt), of the residual magnetic field on the cathode (B_cath), and of the minimum electron-beam radius (r_inj) on the electron beam current (I_el).

magnetic-fields to assure that the radial oscillations of the electron beam are minimal. The electron gun was simulated with an electron beam current ranging from 0.5 A to 10.0 A. Figure 6 shows the dependence of the radial oscillations on the injection magnetic field for electron beam currents 1.0 A and 10.0 A. The injection magnetic field is referred to a peak value in an axial magnetic field distribution near the magnetic shield. For optimum injection, it appears to be approximately 10% higher than the Brillouin magnetic field for each electron current. The amplitude of radial oscillations is measured on the first oscillation of the beam’s envelope as a difference between the maximum radius of the beam and the projected minimum radius. We estimate an error in determining the amplitude of radial beam’s oscillations dr of ±10%. Evidently, the amplitude of radial oscillations is sensitive to the injection magnetic field. A minimum value of the radial beam’s oscillations for optimized geometry and magnetic field is drmin ≈ 0.03 mm for each simulated electron current; this value appears to be independent of the electron current within the ±10% error margins.

Rev. Sci. Instrum. 84, 033303 (2013)

FIG. 9. Electric field distribution in a cathode-anode gap of the electron gun. Values of the electric field are given in V/m.

Figure 7 illustrates the dependence of a simulated electron-beam current on the optimum injection magnetic field and on the electron beam’s radius. With the cathode radius rc = 10.0 mm, and the minimum beam radius rb = 0.8 mm, the area compression of the electron beam is 156 for electron current Iel = 10.0 A. We note that the equilibrium radii of the electron beam for electron currents of 0.5 A and 10.0 A only different by 24%, while the magnetic field changes by almost 50% in this range of electron current. Curiously, the Brillouin theory predicts higher current density in the ion trap for lower electron currents because, with optimum matching, the residual magnetic field on the cathode is smaller. For our electron gun we are planning to use a dispenser cathode and according to the recommendation of the manufacturer its maximum emission current density for long pulses (hundreds of milliseconds) and DC operation should not exceed jemit = 5 A/cm2 . Figure 8 shows the simulated radial distribution of the emission current density from the cathode. The maximum emission current density on the periphery of the cathode is lower than 5 A/cm2 ; hence, we can operate the cathode with nominal heating power and the cathode is expected to have an adequate life time of approximately 2000 h.16 A map of the electric field in a cathode-anode gap for anode voltage Ua = 49.2 kV that is needed for producing electron beam current of Iel = 10.0 A, is shown in Fig. 9. With anode voltage Ua = 49.2 kV the maximum value of the electric field on a negative electrode (Wehnelt) is Emax = 9.8 kV/mm. For these gun parameters, the Kilpatrick17 safety factor for the voltage hold-off is 1.24 and the Latham18 safety factor is 1.66. IV. SUMMARY

FIG. 8. Simulated radial distribution of the cathode’s emission current density for Iel = 10.0 A.

1. We optimized the geometry of the electron gun with electrostatic compression for BNL’s EBIS. Its magnetic structure, with a separate magnet coil in a gun region allows generation of a laminar electron beam in a range of electron current (0.5–10 A) with minimum amplitude of radial oscillations close to 4%.


033303-5


2. The requirements for accurate axial positioning of the gun elements can be met using conventional technologies. 3. The maximum emission current density from the cathode’s surface does not exceed 5 A/cm2 for a 10 A electron beam current and we expect to have a sufficient cathode lifetime. 4. The distribution of the electric field on the gun’s electrodes allows us to operate the gun with acceptable safety factor for electron beam current of at least Iel = 10 A. ACKNOWLEDGMENTS

This work was supported under the auspices of the (U.S.) Department of Energy and the National Aeronautics and Space Administration. 1 J.

Alessi, E. Beebe, S. Binello, L. Hoff, K. Kondo, R. Lambiase, V. LoDestro, M. Mapes, A. McNerney, J. Morris, M. Okamura, A. I. Pikin, D. Raparia, J. Ritter, L. Smart, L. Snydstrup, M. Wilinski, and A. Zaltsman, LINAC2010, FR103 (2010). 2 J. Alessi, E. Beebe, A. I. Pikin, and D. Raparia, in Proceedings of the 2012 Linear Accelerator Conference, Tel-Aviv, Israel, 9–14 September 2012, 2443-TUPB105 (JACoW, 2012). 3 K. Prelec, J. Alessi, and A. Hershcovitch, EPAC1994 (1994), pp. 1435– 1437.

Rev. Sci. Instrum. 84, 033303 (2013) 4 J. G. Alessi, E. Beebe, A. Kponou, A. Pikin, K. Prelec, D. Raparia, J. Ritter,

and S. Y Zhang, AIP Conf. Proc. 572, 196–214 (2001). Kponou, E. Beebe, A. Pikin, G. Kuznetsov, M. Batazova, and M. Tiunov, Rev. Sci. Instrum. 69(2), 1120–1122 (1998). 6 E. N. Beebe, J. G. Alessi, D. Graham, A. Kponou, A. Pikin, K. Prelec, J. Ritter, and V. Zajic, J. Phys.: Conf. Ser. 2, 164–173 (2004). 7 A. Pikin, J. Alessi, E. Beebe, A. Kponou, K. Prelec, G. Kuznetsov, and M. Tiunov, AIP Conf. Proc. 572, 3–19 (2001). 8 A. Pikin, BNL Tech note C-A/AP #245, (2006). 9 J. Arianer, A. Gabrespine, and C. Goldstein, Nucl. Instrum. Methods Phys. Res. 193, 401–413 (1982). 10 J. Faure, P. Antoine, J. C. Ciret, L. Degueurce, P. Gros, A. Courtois, B. Gastineau, R. Gobin, P. Leaux, P. A. Leroy, and J. P. Penicaud, AIP Conf. Proc. 188, 102–113 (1989). 11 M. A. Levine, R. E. Marrs, C. L. Bennett, J. R. Henderson, and D. A. Knapp, AIP Conf. Proc. 188, 82–101 (1989). 12 Z. Xikai, J. Dikui, G. Panlin, S. Shugang, Y. Heping, G. Peirong, W. Naxiu, S. Weiguo, C. Yonglin, X. Xiangyi, F. Shuqing, and Z. Tuantuan, J. Phys.: Conf. Ser. 2, 65–74 (2004). 13 M. A. Levine, R. E. Marrs, J. N. Bardsley, P. Beisdorfer, C. L. Bennett, M. H. Chen, T. Cowan, D. Dietrich, J. R. Henderson, D. A. Knapp, A. Osterheld, B. M. Penetrante, M. B. Schneider, and J. H. Scofield, Nucl. Instrum. Methods Phys. Res. B 43, 431 (1989). 14 F. J. Currell, J. Asada, K. Ishii, A. Minoh, K. Motohashi, N. Nakamura, K. Nishizawa, S. Ohtani, K. Okazaki, M. Sakurai, H. Shiraishi, S. Tsurubuchi, and H. Watanabe, J. Phys. Soc. Jpn. 65(10), 3186 (1996). 15 See http://www.fieldp.com/ for Finite-element Software for Electromagnetics. 16 L. R. Falce, Tech. Dig. - Int. Electron Devices Meet. 29, 448–451 (1983). 17 W. D. Kilpatrik, Rev. Sci. Instrum. 28(10), 824 (1957). 18 High Voltage Vacuum Insulation: Basic Concepts and Technological Practice, edited by R. V. Latham (Academic, London, 1995), pp. 428. 5 A.
