1. Improvement of Beam Quality of the JAEA AVF
Cyclotron for Heavy-ion Microbeam Formation
Satoshi KURASHIMA
Takasaki Advanced Radiation Research Institute
Japan Atomic Energy Agency (JAEA)
Thank you Mr. Chairman, I am Satoshi Kurashima from JAEA. This is my first presentation in English, so, please excuse green presentation.
930 AVF Cyclotron
Resonator

2. Members
Cyclotron R & D
S. Okumura, W. Yokota, S. Kurashima, N. Miyawaki, T. Nara,
I Ishibori, K. Yoshida, H. Kashiwagi, Y. Yuri, T. Yuyama, T. Ishizaka
Operators
T. Yoshida, K. Akaiwa, S. Ishiro, T. Yoshida,
S. Kano, A. Ihara, K. Takano, S. Mochizuki
Microbeam R & D
T. Kamiya, M. Oikawa, T. Satoh
This is members concerning this work. Staff and operators of 20 people develop and maintain the cyclotron, and experts of 3 develop a heavy-ion microbeam system.

3. Outline Introduction Microbeam Formation System
Required Developments for High Quality Beam
Development of Flat-top (FT) Acceleration System
Beam Tuning for Reduction of the Energy Spread
Microbeam Formation Experiment
Quick Change of Ion Species of Heavy-ion Microbeam
Summary
This is outline of the presentation. 1 Introduction and 2 I talk about how to make a microbeam with focusing lenses. 3 many developments are required for the cyclotron to make the microbeam. 4 this is my work, and I explain in detail. 5, 6 we tried to form the heavy-ion microbeam with the energy reduced ion beam. 7, and summary.

4. Energy Range Covered by the Four Accelerators
1. Introduction
TIARA Facility
Takasaki Ion Accelerators for Advanced Radiation Application
・K110 AVF Cyclotron
・3 MV Tandem Accelerator
・3 MV Single-ended Accelerator
・400 kV Ion Implanter
Ion Beam
RF Amplifier
Deflector
Magnetic Channel
RF Resonator
Flat-top
Resonator
Dee Electrode
Main Coil
Gradient
Corrector
Phase Probe
Cyclotron Magnet
(Side & Lower Yoke)
Specification of the JAEA AVF cyclotron
Energy Range Covered by the Four Accelerators
・K110 AVF Cyclotron
・3 MV Tandem Accelerator
・3 MV Single-ended Accelerator
・400 kV Ion Implanter
An ion beam irradiation facility TIARA (Takasaki Ion Accelerators for Advanced Radiation Application) was established in 1993 to apply ion beams for research in biotechnology and materials science. TIARA has four accelerators, an AVF cyclotron, a 3 MV tandem accelerator, a 3 MV single-ended accelerator and a 400 kV ion implanter. The four accelerators cover a wide range of ion beam energy and mass number.
This figure shows main components of the JAEA AVF cyclotron and specification of the cyclotron are here. Bending limit of 110, focusing limit of 95, Extraction radius of m, maximum average field of 1.64 T, RF 11 to 22 MHz, and acceleration harmonics of 1, 2, and 3 are available. The cyclotron can accelerate protons from 10 to 90 MeV, and heavy ions from 2.5 to 27 MeV/n.

5. Microbeam by a collimation aperture. Beam size of 5-10 um.
Hyper Nanogan
14.5 GHz,
Metal Ions by MIVOC
Si, Fe, Ru, Au
Octopus
6.4 GHz and 14.3 GHz,
The oldest ECRIS
operating in the world ?
Microbeam by a collimation aperture.
Beam size of 5-10 um.
A few ions per minute. 10 μ m 核
イオンの
ヒット位置
Hit position
Cell nucleus
This figure shows a layout of instruments in the cyclotron facility. We have twelve horizontal irradiation ports and four vertical ports. Three ion sources are available, multi-cusp for proton and deuteron, very old ECR ion source “Octopus” for heavy-ion and new ECR ion source “Hyper-Nanogan” for heavy-ion including metal ions by MIVOC. Both of the ECR are not type of superconducting, so it is difficult for us to accelerate ion beams up to GeV.
One of special techniques of TIARA facility is heavy-ion microbeam irradiation to living cells or semiconductor IC chips. (push) This figure shows cross-sectional view of the vertical irradiation ports. At this port, microbeams with a spot size of 5 to 10 um are formed by a collimation aperture for biotechnology research like apoptosis or bystander effect. We can irradiate cells one by one with a movable sample stage. Beam users want to irradiate the cell nucleus precisely, but it’s difficult because the beam size is comparable to or larger than the cell nucleus. Moreover, hitting rate is very slow.
So we started to develop a new microbeam system which can make the microbeam with a spot size of 1 um by a focusing lens system and can irradiate cells very fast.
Microbeam by a focusing lens system.
Beam size of 1um.
600 ions per minute.

6. 2. Microbeam Formation System
5 um
1 um
Challenging Development for the Cyclotron
This figure shows a drawing of the new microbeam system. This system mainly consists of micro-slits, divergence defining slits and quadruplet quadrupole magnets. We can target and hit by this beam scanner quickly. However, it is so easy for us to focus the ion beam to 1 um by this system because of chromatic aberration in the lens.
We estimated the beam size for several conditions by the TRANSPORT code. This figure shows calculation results for the case of energy spread dE/E = 0.1%; it’s very good condition for cyclotrons. However, the microbeam cannot be focused to 1 um even if the divergence angle is restricted to 0.2 mrad. On the other hand for the case of dE/E = 0.02%, we can focus the ion beam to 1 um, and can focus about 1 um for large divergence angle. Since beam current of the heavy-ion with high charge state is not so high, we don’t want to close the divergence slits if possible.
Reduction of beam energy spread under the order of 10-3 is required for the cyclotron, it’s very challenging development. We decided to develop a flat-top acceleration system to provide a high quality beam.
Flat-top Acceleration
Beam optics

7. 3. Required Developments for High Quality Beam
Many developments are necessary for acceleration of the high quality beam. We selected fifth-harmonic frequency for the FT acceleration. The FT acceleration with the fifth-harmonic frequency has advantages, compact size of an additional resonator, low voltage of 4% of the fundamental dee voltage, and so on. But a wide range of resonance frequency is required from 55 to 110 MHz. I talk about the development of the FT system later.
Next is stability of cyclotron magnetic field dB/B smaller than 0.002%. dB/B of the order of 10-6 was achieved by controlling temperature of the cyclotron magnet yoke.
Control of beam phase width within 10 degrees.., Dr. Fukuda will talk about this point in deal at the next presentation.
High stabilities of the acceleration voltage are needed. We improved them by precise control of temperature around feedback circuit within degrees.
The beam current is reduced by phase slits. In order to cover the beam reduction, a high performance beam buncher using saw-tooth waveform voltage was developed.

8. 4. Development of Flat-top Acceleration System
The cyclotron accelerates a wide range of ion species and energy. The goal of this work is to accelerate all ion beams available at our facility with the FT acceleration. So, we had many tasks.
This figure shows an outline of the cyclotron resonator. The fundamental frequency is changed 11 to 22 MHz by sliding the movable short. The FT resonator is coupled here; a port for a cryogenic pump is used. The FT resonator was designed by the MAFIA code. The fifth-harmonic frequency is determined by both the capacitance of a coupling electrode C5 and position of the movable-short L5. From the result of simulation, specification of the FT resonator was determined, and the resonator was installed to the cyclotron.
This figure shows relation between the resonance frequency and the parameters of the FT resonator in a cold power test. For higher frequency, we set the gap of C5 wide, and set the position of L5 short. Required range of the resonance frequency is fully covered by the FT resonator.

9. Power Test
This figure is a photograph of the resonators. This is the main resonator and this is the FT resonator, small resonator.
Next step, we tried a high power test. This figure shows “flat-top” waveform observed with the dee voltage pick-up. The fundamental frequency and voltage are MHz and 25 kV, respectively. Only one dee voltage pick-up is set at the dee electrode, so this waveform is divided into two frequencies by a filter circuit. As a result, we succeeded observing the FT waveform from 59 to 102 MHz.

10. Preparation for the FT acceleration
Key points
・Isochronous field tuning within ±5deg.
・Precise control of the beam phase width by phase slits.
・fine-tuning of trim coil current at the center region.
・Estimation of the optimum voltage ratio of the fifth-harmonic frequency to the fundamental one.
Very careful tuning of the cyclotron is needed before operating the FT resonator. First is isochronous field tuning within +-5 degrees. We can tune the isochronous field precisely using a phase probe which has ten pairs of pick-up electrode. As a recent trend, tuning within degrees is easily achieved.
Precise control of the beam phase width by phase slits will be presented by Dr. Fukuda.
Fine-tuning of trim coil at the center region. Twelve pairs of trim coils are installed. We fine-tune the smallest #1 and #2 trim coil in order to set the beam bunch around the top of the sinusoidal waveform.
Estimation of the optimum voltage ratio of the fifth-harmonic frequency to the fundamental one. When the voltage distribution of the fifth-harmonic frequency along the dee electrode is uniform, we set the voltage 1/25 (one-twenty fifth) of the fundamental voltage.
However, the cyclotron has clear voltage distribution as shown in this figure. These data were measured by vector-volt meters. The voltage drops significantly for higher frequency. On the other hand, negligible voltage distributions were measured for the fundamental frequency.
We estimated the voltage ratio of the fundamental frequency to the fifth-harmonic one at the tip of the dee electrode to compensate the voltage distribution as shown in this figure. Since the span angle of the dee is not 90 degrees (it’s 86 degrees), differences occurred according to the acceleration harmonics.

11. 5. Beam Tuning for Reduction of the Energy Spread
Deflector electrode
Deflector
probe
Moving direction
Beam
bunch
probe head
The turn separation can be seen by the scanning current probe with thin probe head when beam width is reduced by the FT acceleration.
Ion Beam: 260 MeV 20Ne7+
N = 265 (h=2)
We started beam development of 260 MeV 20Ne7+ with acceleration harmonics of 2. The cyclotron takes out the ion beam by multi-turn extraction and we could not observe turn separation. But, single-turn extraction is indispensable to provide a high quality beam. In order to confirm the single-turn extraction easily, we developed new deflector probe with a thin Mo sheet, thickness of 0.5 mm, as a probe head. Turn separation can be seen by the deflector probe when the beam width is reduced by the FT acceleration.
As a result, very clear turn separation was observed. The ratio of the acceleration voltages is close to the calculation result. Extraction efficiency is improved from 60% to 95%.

12. Energy Spread Measurement
The energy spread DE/E was measured by analyzing magnet with a micro-slit system.
We estimated the energy spread dE/E of the Ne beam from the beam spread caused in the analyzing magnet. In order to improve the resolution of the system, width of injection beam was defined at 0.1 mm with a micro-slit. This figures show the results. Energy spread dE/E was estimated to be 0.1% in FWHM for FT off, and 0.05% for FT on. Energy spread of the beam was reduced with the FT acceleration system.

13. 6. Heavy-ion Microbeam Formation
5 x 5 Single-ion Hit Pattern
1 um microbeam
10 m
(a)
Optical Microscope Image
Finally, we tried to form a heavy-ion microbeam with the Ne beam. Usually, size of the microbeam is estimated by analyzing a SEM (secondary electron microscope) image of a copper fine mesh. This is an optical microscope image of the copper mesh used in this estimation. Distance of each line is 25 um.
After tuning of the micro-slits, divergence defining slits and quadrupole magnets, a very clear SEM image was observed. The beam size was estimated to be about 0.7 um. This is the highest energy as heavy-ion sub-micron beam in the world.
We hit a plastic plate CR-39 controlling number of ion by pulse-chopping system installed at the injection line, and controlling the target point by the beam scanner. This is the result after chemical etching. Clear 5 x 5 pattern can be seen. Single-ion hit with targeting accuracy of 1 um was confirmed.

14. 7. Quick Change of Ion Species of Heavy-ion Microbeam
100 to 2000 keV/um
in water.
260 MeV Ne
Wide range of linear energy transfer (LET) is required for research in biotechnology and materials science.
Cocktail beam acceleration
Quick change of ion species of heavy-ion microbeam. Somehow, we succeeded to form the microbeam of Ne. Users of the microbeam require a wide range of linear energy transfer (LET), moreover they hope quick change in one beam time. We have to do beam developing for various ion species. It must take a lot of time… But, accelerator we use is cyclotron. We can select the cocktail beam acceleration technique to change the ion species quickly.
We selected ion species with M/Q near 2.8 that is one of 20Ne7+. We can accelerate these ions by slightly changing the rf frequency corresponding to the difference of M/Q. If we can accelerate all ion species with the FT system,
a series of microbeam with very wide range of LET is available in a beam time. This is the cocktail microbeam formation. However, production of 56Fe20+ is hard for our ECRIS. We have to develop a superconducting ECRIS in the near future. At the beginning, we tried the cocktail microbeam formation with Ne and Ar.
Cocktail Microbeam Formation

15. Cocktail Microbeam Formation
SEM Image by Ne
SEM Image by Ar
・Optimization of ECRIS for each ion.
・Fine-tuning of trim coil of center region and harmonic coils.
It is not difficult to accelerate Ar beam, but high stability of the beam current is required during the microbeam tuning. We have to optimize the condition of the ECRIS for each ion. For cocktail beam acceleration, the beam trajectories are almost the same. So, we fine-tune central trim coil and harmonic coil for clear turn separation and better extraction efficiency.
This figure shows deflector probe patterns of Ne and Ar. A little difference is observed, but extraction efficiencies are almost the same. No need for re-tuning in the beam transport line because the ion beams have the same magnetic rigidity. However, fine-tuning was required to just meet the beam line of the microbeam system.
This is SEM image of Ne, standard condition. And next, SEM image of Ar. Very clear images were observed. It took about 30 minutes to change Ne microbeam to Ar one. I think the cocktail microbeam formation is an epoch technique in the field of microbeam research.
Quick change within 30 minutes.
(Usually, 6 hours)
Very powerful tool in microbeam research!

16. Summary
The FT acceleration system with fifth-harmonic frequency improved the beam quality and extraction efficiency.
Heavy-ion microbeam with a spot size and hitting accuracy of 1 um was successfully formed.
Ion species of the microbeam was quickly changed by the cocktail beam technique within 30 minute.

17. Appendix. Beam Transmission
Sinusoidal + Saw-tooth buncher
1.4 times beam intensity
Appendix. This figure shows typical beam transmission efficiency for each acceleration harmonics. The standard point is IS1 where is just downstream of the ion source. Inflector, before and after deflector and out of the cyclotron. For h = 1 and 3, about 10% is typical value. Usually, harmonics number of 2 results good transmission over 15%.
On the other hand, the transmission for the FT acceleration is 10%, not so low. We cut the injection beam by slits at IS5 to improve the emittance. However, simultaneous operation of the sinusoidal and saw-tooth buncher recovers the beam current. As a result, the beam current enough for the microbeam tuning is obtained.
Injection Line
Deflector

18. Thank you very much for your attention !