Embed Size (px)
Citation preview
Investigations of the emittance and brightness of ion beams from an electron beam ionsource of the Dresden EBIS typeAlexandra Silze, Erik Ritter, Günter Zschornack, Andreas Schwan, and Falk Ullmann Citation: Review of Scientific Instruments 81, 023303 (2010); doi: 10.1063/1.3284512 View online: http://dx.doi.org/10.1063/1.3284512 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/81/2?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Neutralization of space charge on high-current low-energy ion beam by low-energy electrons supplied fromsilicon based field emitter arrays AIP Conf. Proc. 1496, 368 (2012); 10.1063/1.4766565 Compact microwave ion source for industrial applicationsa) Rev. Sci. Instrum. 83, 02B914 (2012); 10.1063/1.3673864 Production of low-Z ions in the Dresden superconducting electron ion beam source for medical particle therapya) Rev. Sci. Instrum. 83, 02A507 (2012); 10.1063/1.3672110 A 2.45 GHz electron cyclotron resonance proton ion source and a dual-lens low energy beam transporta) Rev. Sci. Instrum. 83, 02A329 (2012); 10.1063/1.3669802 High-brightness source for ion and electron beams (invited) Rev. Sci. Instrum. 69, 1026 (1998); 10.1063/1.1148532
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 03:48:32
Investigations of the emittance and brightness of ion beamsfrom an electron beam ion source of the Dresden EBIS type
Alexandra Silze,1,a� Erik Ritter,1 Günter Zschornack,1 Andreas Schwan,2 andFalk Ullmann2
1Institut für Angewandte Physik, Technische Universität Dresden, Helmholtzstr. 10,D-01062 Dresden, Germany2DREEBIT GmbH, Zur Wetterwarte 50, D-01109 Dresden, Germany
�Received 4 August 2009; accepted 12 December 2009; published online 4 February 2010�
We have characterized ion beams extracted from the Dresden EBIS-A, a compact room-temperatureelectron beam ion source �EBIS� with a permanent magnet system for electron beam compression,using a pepper-pot emittance meter. The EBIS-A is the precursor to the Dresden EBIS-SC in whichthe permanent magnets have been replaced by superconducting solenoids for the use of the sourcein high-ion-current applications such as heavy-ion cancer therapy. Beam emittance and brightnessvalues were calculated from data sets acquired for a variety of source parameters, in leaky as wellas pulsed ion extraction mode. With box shaped pulses of C4+ ions at an energy of 39 keV root meansquare emittances of 1–4 mm mrad and a brightness of 10 nA mm−2 mrad−2 were achieved. Theresults meet the expectations for high quality ion beams generated by an electron beam ionsource. © 2010 American Institute of Physics. �doi:10.1063/1.3284512�
I. INTRODUCTION
Collisions of ions and atoms in various ways are amongthe most fundamental processes in the universe. Thus, beamsof highly charged ions �HCIs� have become important toolsin a wide range of basic and applied research activities. Theyare used for investigations of atomic processes in laboratoryand astrophysical plasmas,1 nuclear physics experimentswith radioactive beams accelerated to energies beyond thecoulomb barrier,2 as well as in materials science for surfacemodification and analysis.3,4 Another growing field in whichHCI beams are used is cancer therapy. For several types oftumors carbon ion radiation therapy has proven to be ahighly advantageous alternative to photon radiation therapy.5
However, in many cases efficient ion beam transportation orprecise positioning of the ion impact region on a target isrequired. Therefore, the applied HCI beams need to be oflow emittance and high brightness.
Electron beam ion sources or traps �EBIS/T� are wellsuited for tasks with high requirements on the ion beam qual-ity because the region in which ions are produced is smalland the temperature of the created ions is low. They havealready replaced more traditional ion sources such as elec-tron cyclotron resonance �ECR� and liquid metal ion sources�LMISs� in several fields, such as materials research.6–8 Onthe other hand, focused ion beam setups, ion beam lithogra-phy, or cancer therapy facilities are currently still relying onLMIS and ECR ion source technology due to the compara-tively low ion currents delivered by EBIS/T systems. Never-theless, with an increased ion output the use of EBIS/Twould have many advantages also in these fields, as dis-cussed in previous publications.9,10
Improvement in EBIS/T technology has made thesesources more and more feasible for high-current applications.Recently, the Dresden EBIS-SC has been developed whichcontinues a series of HCI sources with a unique concept interms of tabletop size and resource efficiency.11 Featuring anenhanced electron current density and trap capacity theEBIS-SC is designed to meet the required ion numbers perpulse for cancer therapy. However, it is only able to competewith ECR sources in this field if the beam quality is suffi-cient enough to avoid losses during the injection of ionpulses into the accelerator part of the irradiation facilities.
A quantitative measure of the quality of ion beams isgiven by the root-mean-square emittance
�x,rms = �x2 · x�2 − xx�2. �1�
Herein, x2, x�2, and xx�2 are the moments of particle distri-bution in x-x� trace space12 �equivalent in case of y-y��. Acommon method to determine these ion beam parameters isthe pepper-pot technique which can be traced back to worksof the 1960s.13 It involves the use of a mask of holes cuttingout “beamlets” from the investigated ion beam. The beamletsare projected onto a screen in the distance L behind themask. The x- and y-axis projections of the created picturecan then be analyzed revealing the desired information. Forthe emittance calculation in the x-dimension, the centralbeamlet positions on the screen, Xj, have to be recorded.Together with the original hole positions, xj, they define themean directions of the beamlets, xj�= �Xj −xj� /L. The spreadof the individual beamlets also has to be taken into accountin the form of the standard deviations of their Gaussian-likeprojections, � j. With the areas under the peaks, nj, the overallsum of counts N=� jnj, as well as the mean values x̄ and x�one can then calculate the momentsa�Electronic mail: [email protected].
REVIEW OF SCIENTIFIC INSTRUMENTS 81, 023303 �2010�
0034-6748/2010/81�2�/023303/5/$30.00 © 2010 American Institute of Physics81, 023303-1
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 03:48:32
x2 �1
N�
j
nj�xj − x̄�2, �2�
x�2 �1
N�
j
nj��� j/L�2 + nj�xj� − x��2� , �3�
and
xx� �1
N�
j
nj�xj − x̄��xj� − x�� , �4�
defining the rms-emittance as presented in formula �1�. For amore detailed description of the calculation of the rms-emittance with a pepper-pot system, see also Ref. 14.
In this paper, we present rms-emittance and brightnessmeasurements on the precursor to the Dresden EBIS-SC, theDresden EBIS-A, using a pepper-pot emittance meter. Addi-tionally, trace space distributions referring to the calculatedemittances are shown which provides further informationabout the condition of the investigated beam.
II. EXPERIMENTAL SETUP
Electron beam ion sources and traps generally consist ofa sequence of rotationally symmetric electrodes including asmall, heated cathode of high electron emissivity, typicallyfollowed by several collinear drift tubes, an electron collec-tor, and an ion extraction electrode. The drift tube ensembleof the Dresden EBIS-A is comprised of three elements andits extraction channel includes an additional einzel lens forimproving the shape of the delivered beam. Along the centerof the drift tube alignment a linear electron beam is formed,which enables subsequent electron impact ionization of theatoms inside the source. Ion trapping in the radial direction isrealized by the space charge potential of the electron beam.In the axial direction the ions are captured by the box-shapedpotential produced by three different voltages on the drifttubes. The maximum achievable charge state within a certainionization time and in general the speed with which the ion-ization takes place depends on the energy of the electronbeam and its current density. For higher ionization rates theelectron beam is compressed by a magnetic field created bymagnets in Helmholtz configuration. In case of the DresdenEBIS-A permanent magnets and auxiliary soft iron compo-nents are used to produce a flattop magnetic flux distributionof 620 mT along the electron beam axis. With this setupelectron densities of several hundred A /cm2 can beachieved.15
For the experiments presented in this paper we con-nected the Dresden EBIS-A to a beamline featuring severalion optical and beam guiding elements, as well as a 90°bending magnet for charge state selected ion beam analysis.The complete setup is shown in Fig. 1. For a first series oftests, the pepper-pot emittance meter was mounted behindthe einzel lens downstream from the source to gain informa-tion about the unseparated beam. Afterwards, charge sepa-rated carbon ion beams were investigated behind the bendingmagnet.
Next to the positions of the pepper-pot emittance meterduring the measurement, a schematic picture of the pepper-
pot setup itself is given in Fig. 1. The design is similar to adevelopment by Pikin and co-workers.16 The pepper-potmask of uniformly distributed holes with a distance of 1 mmand a diameter of 0.2 mm each is mounted 54.0�0.5 mmaway from a chevron-type multichannel plate �MCP�. TheMCP is followed by a phosphor screen to visualize the sig-nal. With the help of a mirror the screen is imaged onto acharge coupled device �CCD� camera placed outside of thevacuum channel. The pepper-pot emittance meter is mountedon a linear feedthrough. This allows for moving the setup inand out of the beamline center enabling further investiga-tions on the ion beam, while the pepper-pot is installed in thefacility.
III. MEASUREMENT OF THE EMITTANCE OF IONBEAMS PROVIDED BY THE DRESDEN EBIS-A
To measure the trace space distribution and the emit-tance of ion beams the pixel scale of the images takenwith the pepper-pot camera first had to be calibrated.Therefore, pictures of the screen area were taken whichfeatured parts of the setup with known dimensions. Theconversion factor from pixels to millimeter resulted in0.0550�0.0006 mm /pixel.
The first emittance measurements took place at pepper-pot position 1 in the beamline, behind an einzel lens down-stream from the source. The cathode was set to a negativevoltage of �3.0 kV, while the drift tube voltage was U0
=5.7 kV. The electron beam therefore had an energy of 8.7keV. The electron current was set to Ie=60 mA. For a con-tinuous extraction of ions from the EBIS an asymmetric axialtrap was applied allowing the ions to escape toward thebeamline. The difference in the voltages applied to the cen-tral and last drift tube was Ut=20 V, which represents thetrap depth or the potential well the ions had to overcome.The pressure inside the EBIS was set to 5�10−9 mbar bycontinuous propane gas injection. Propane was chosen forthe production of highly charged carbon ions because of theirhigh relevance for applications such as cancer therapy.
With these parameters the pepper-pot image presented inFig. 2 was recorded. The trace space distributions of thebeam in x- and y-direction are displayed in Fig. 3. The el-
90° Bendingmagnet
DresdenEBIS-A
Einzellens
CCDcamera
Position 1
Position 2
PepperpotmaskPhosphor screen + MCP
Mirror
Faradaycup
FIG. 1. �Color online� Overview of the experimental setup. The beamline,the pepper-pot setup, and its positions in the beamline during the experi-ments.
023303-2 Silze et al. Rev. Sci. Instrum. 81, 023303 �2010�
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 03:48:32
lipses fitted to the distributions represent the four-times rmsemittances, i.e., trace space areas four times as large as thearea of the calculated rms emittance, which practically in-clude close to 100% of the investigated beam.12 The ellipsesare almost perfectly upright for both dimensions. Therefore,it can be concluded that the investigated beam with a diam-eter of about 13 mm was relatively parallel. The rms-emittances of the unseparated beam with an energy of Ei
=q ·U0 resulted in �x,rms=6.1�3.3 mm mrad and �y,rms
=6.9�2.5 mm mrad. The errors of the emittance values in-clude the uncertainties of both the distance between maskand MCP, as well as the pixel to millimeter conversion. Fur-thermore, the statistical errors from the fit of the peaks withmultiple Gaussians are taken into account. The largest con-tribution, about 60% of the error, comes from the finite sizeof the pepper-pot holes. It affects the width of the beamletprojections systematically and causes the emittance values toappear larger than they are in reality. Therefore, it should benoted that, within the given error bars, the actual ion beamemittance is more likely to be lower and less likely to behigher than the presented values.
A factor which influenced the results and has not beentaken into account in the error discussion is the MCP voltage.However, it was carefully chosen to give the best possiblecontrast between the highest peak intensities at the centerand lowest peak intensities on the side of the projection ofthe beamlets. The MCP was operated about 100 V below thevoltage where saturation effects started influencing the inten-sity distribution and the resulting emittance value signifi-cantly. On the other hand, it was set high enough to allow fora visualization of the full width of the beam. In case of thepicture presented in Fig. 2 the applied voltage had a typical
value of UMCP=800 V. Within a range of about �UMCP
= �100 V the calculated emittance stayed well within thegiven error bars. For the following experiments, dependingon the change in beam intensity, the MCP voltage was ad-justed accordingly.
To determine how sensitive the beam properties are tochanges in the ion optics, the trace space distributions andrms-emittances for different voltages applied to the einzellens in the beamline were measured. Figure 4 shows how theshape and the angle of the beam’s trace space ellipse changeswith increasing lens potential, Uel. As the beam convergesmore strongly the ellipse becomes subsequently more tiltedclockwise. At Uel=3.2 kV the focus of the beam reached thepepper-pot position. With the voltage on einzel lens 2 ex-ceeding this value the semimajor axis of the ellipse crossedthe y-axis indicating that the beam diverges again at the po-sition of the pepper-pot system. The mean emittances duringthis measurement were �x,rms=5.9 mm mrad and �y,rms
=6.3 mm mrad with standard deviations of 1.0 and 1.5mm mrad, respectively.
After the first tests at position 1 had been concluded, thepepper-pot setup was transferred to position 2 where beamsof ions with different q /A ratios were investigated individu-ally. Figure 5 shows a pepper-pot image obtained for a con-tinuous beam of C4+ ions extracted from the source operatedat the same parameters as described before. The x-axis pro-jection of the plot reveals two C4+ beam components whichhad been slightly separated by the magnet. This, again, is aneffect which is related to the different electric potentials inthe drift tube region where the ions are created. However, theweak signal on the left side of the picture is produced by lessthan a 10% fraction of the beam. Neglecting this componentthe emittances of the main C4+ beam with an energy ofEC4+=23 keV resulted in �x,rms=3.5�1.7 mm mrad and�y,rms=2.8�1.5 mm mrad.
The electron current of an EBIS limits the amount ofcaptured ions inside the source and influences the tempera-ture of the produced plasma in the trap. Thus, it is expectedto be significantly relevant for the emittance of the extractedion beam. In fact, as presented in the upper graph in Fig. 6,the x- and y-emittances increase with higher electron cur-
60
40
20
8006004002000
a.u
.
X (pixel)
60
04
00
20
00
a.u.Y
(pix
el)
40
20
60 0
FIG. 2. �Color online� Pepper-pot image and projections on the x- andy-axis. For the calculation of the emittance the projections were fitted withmultiple Gaussians.
4
0
-4
x'(m
rad
)
840-4x (mm)
4
0
-4
y'(m
rad
)
-8 -4 0 4y (mm)
FIG. 3. �Color online� Trace space plots of the x- and y-dimension of thebeam imaged in Fig. 2.
x'(mrad)
x (mm)-6 -4 -2 0 2 4 6
2000 V1500 V
1000 V
2500 V3000 V
3500 V
0
8
16
24
-8
-16
-24
FIG. 4. Illustration of the change of the trace space ellipses for differenteinzel lens voltages Uel.
023303-3 Silze et al. Rev. Sci. Instrum. 81, 023303 �2010�
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 03:48:32
rents, Ie. Mainly, this is related to a higher number ofproduced C4+ ions, which is indicated by the graphbelow showing the change in the extracted C4+ current. Toillustrate in which way the trace space density of the beamis influenced by the variation of Ie, the brightness, Brms
= IIon / ��2�x,rms�y,rms�, is also given in the graph. In the rangebetween 0.1 and 0.2 pA mm−2 mrad−2, it rises with higherelectron currents since the relative increase of the extractedion current is larger than the relative increase of the productof emittances. However, the brightness gain at higher elec-tron currents is smaller than at lower currents. Therefore,depending on the demands of a specific application, it isoften useful to work with high electron currents but it is notnecessarily expedient to run the source at its limit.
For the final series of measurements the operation pa-rameters of the facility were adjusted to approach require-ments occurring in carbon ion radiation therapy. The drift
tube voltage was set to U0=9.8 keV, the electron currentwas raised to 90 mA, and a trap voltage of Ut=100 V wasapplied. After the C4+ current on Faraday cup 2 had beenmaximized with the help of the ion optics, the operationmode of the source was changed to pulsed extraction with acycle time of 30 ms. Furthermore, the slope of the decreasingtrap voltage, Ut, during the process of opening the trap wasredesigned to allow for an extraction of plateau-shapedpulses. This is important for applications where stable con-ditions during the irradiation need to be assured. With a rateof 30 frames per second pepper-pot images of individual C4+
ion bunches with an energy of EC4+=39 keV were taken. InFig. 7 an exemplary pepper-pot image for this measurementis displayed. As opposed to the leaky mode experimentswhere a weak second beam component was visible, herein,the signal on the screen was produced exclusively by ionsfrom the central drift tube area. The average emittances of 10pulses resulted in �x,rms=4.1�2.4 mm mrad and �y,rms
=1.4�0.8 mm mrad. The associated standard deviations forthis series of pulses were less than five percent in each di-mension and the differences in the trace space areas occupiedby the individual ion bunches were marginal. Within a dura-tion of 50 �s the C4+ pulses reached an ion current of IC4+
�600 nA. Together with the measured emittance values thisresults in an achieved ion beam brightness of Brms
�10 nA mm−2 mrad−2 in pulsed mode operation of theEBIS.
IV. CONCLUSION AND OUTLOOK
With the help of a pepper-pot emittance meter we haveaccomplished to gain detailed information about the ionbeam extracted from a room temperature electron beam ionsource of the Dresden EBIS/T type. We have carried outemittance and brightness measurements with a carbon ionbeam before and after charge state separation, at varioussource parameters, as well as different operation modes ofthe EBIS. The values of the emittance were on the order of1–10 mm mrad. In leaky mode, the brightness of the beamvaried between 0.1 and 0.2 pA mm−2 mrad−2, while inpulsed mode a value of 10 nA mm−2 mrad−2 was reached.
For carbon ion radiation therapy ECR ion sources areapplied delivering C4+ currents of several hundredmicroamps.17 In comparison with this the currents extracted
0 200 400 600 800
6
0
-640-4
x (mm)
x’(mrad)
6
0
-6
40-4y (mm)
y’(mrad)
x (pixel)
FIG. 5. �Color online� Pepper-pot image and trace space plots of a beam ofC4+ ions at EC4+=23 keV behind the bending magnet. The x-axis projectionshows a separation of two beam components within the C4+ beam. Thestronger signal on the right side is produced by ions originating from thecenter of the axial trap of the EBIS. The weak signal on the left refers to asmall amount of ions which are created in the first drift tube section and startfrom a slightly higher electrostatic potential.
2
4
50 60 70Ie (mA)
0
20
I C4+(pA)
Brms
(pAmm-2mrad-2 )
0.2
0.050 60 70
Ie (mA)
0
6
2
4
0
6
FIG. 6. Top: Results of the C4+ beam emittance for different electron cur-rents Ie. Bottom: Developing of the ion current IC4+ extracted from thesource and the brightness B of the beam during the measurement of thevalues presented in the previous graph �top�.
6
0
-6
40-4x (mm)
x’(mrad)
4
0
-4
-3 0 3
60
y (mm)
y’(mrad)
0 200 400 600 800x (pixel)
FIG. 7. �Color online� Pepper-pot picture and trace space plots of the C4+
pulse at EC4+=39 keV.
023303-4 Silze et al. Rev. Sci. Instrum. 81, 023303 �2010�
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 03:48:32
from the Dresden EBIS-A are certainly low. However, thebeam brightness of this source in pulsed mode is an order ofmagnitude better than what has been presented for ECRsources.17 Also, we have managed to extract plateau-shapedion pulses from the EBIS to avoid the necessity of a beamchopper which has to be used in combination with anECR ion source. With the Dresden EBIS-SC a new genera-tion of tabletop EBIS will be available combining the goodbeam quality and resource efficiency achieved in previousmodels with extraction currents high enough to compete withECR ion source in this field of application. The firsttest measurements with the EBIS-SC are currently beingperformed.
ACKNOWLEDGMENTS
We thank our colleagues Martin Kreller and MikeSchmidt, as well as Jochen Pfister for helpful discussions.This work was supported by the EFRE fund of the EU, bythe Freistaat Sachsen �Project Nos. 12321/2000 and 12184/2000�, and by GSI, Darmstadt, Germany �Project No.DD.ZSCH�.
1 F. I. Allen, C. Biedermann, R. Radtke, and G. Fussmann, Phys. Rev. A 78,032705 �2008�.
2 F. Wenander, Nucl. Phys. A 746, 40 �2004�.3 S. Momota, Y. Nojiri, J. Taniguchi, I. Miyamoto, N. Morita, and N. Ka-wasegi, Rev. Sci. Instrum. 79, 02C302 �2008�.
4 T. Schenkel, A. Persaud, A. Kraemer, J. W. McDonald, J. P. Holder, A. V.Hamza, and D. H. Schneider, Rev. Sci. Instrum. 73, 663 �2002�.
5 D. Schulz-Ertner and H. Tsujii, J. Clin. Oncol. 25, 953 �2007�.6 R. Heller, S. Facsko, R. A. Wilhelm, and W. Möller, Phys. Rev. Lett. 101,096102 �2008�.
7 T. Peters, C. Haake, D. Diesing, D. A. Kovacs, A. Golczewski, G.Kowarik, F. Aumayr, A. Wucher, and M. Schleberger, New J. Phys. 10,073019 �2008�.
8 M. Tona, Y. Fujita, C. Yamada, and S. Ohtani, Phys. Rev. B 77, 155427�2008�.
9 M. Schmidt, H. Peng, G. Zschornack, and S. Sykora, Rev. Sci. Instrum.80, 063301 �2009�.
10 R. Becker, Rev. Sci. Instrum. 77, 03A903 �2006�.11 G. Zschornack, M. Kreller, V. P. Ovsyannikov, F. Grossman, U. Kentsch,
M. Schmidt, F. Ullmann, and R. Heller, Rev. Sci. Instrum. 79, 02A703�2008�.
12 M. Reiser, Theory and Design of Charged Particle Beams �Wiley, NewYork, 1994�.
13 L. E. Collins and P. T. Stroud, Nucl. Instrum. Methods 26, 157 �1964�.14 M. Zhang, Emittance Formula for Slits and Pepper-Pot Measurement
�Fermi National Accelerator Laboratory, Batavia, IL, 1996�, Technote.15 A. Silze, G. Zschornack, V. P. Ovsyannikov, and F. Ullmann, Rev. Sci.
Instrum. 79, 083302 �2008�.16 A. Pikin, A. Kponou, J. Ritter, and V. Zajic, Pepper Pot Emittance Meter
�Brookhaven National Laboratory, Upton, NY, 2006�, Technote.17 T. Winkelmann, R. Cee, T. Haberer, B. Naas, A. Peters, S. Scheloske, P.
Spädtke, and K. Tinschert, Rev. Sci. Instrum. 79, 02A331 �2008�.
023303-5 Silze et al. Rev. Sci. Instrum. 81, 023303 �2010�
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
130.70.241.163 On: Sun, 21 Dec 2014 03:48:32
