IMPORTANT NUCLEAR SCIENCE ENGINEERING · Web viewReport of Beijing Radioactive Ion-B. eam. Facilities (BRIF) in 2006. Engineering Division of BRIF Project. Beijing Radioactive Ion-beam

IMPORTANT NUCLEAR SCIENCE ENGINEERING·Beijing Radioactive Ion-Beam Facility 12

Beijing Radioactive Ion-Beam Facility

Progress Report of Beijing Radioactive Ion-BeamFacilities (BRIF) in 2006

Engineering Division of BRIF Project

Beijing Radioactive Ion-beam Facilities (BRIF) project is in progress. The preliminary engineering design of BRIF has been finished. The working drawing design of the project and some critical equipment ordering were started in 2006.

1 CyclotronThe compact 100 MeV, 200 µA proton cyclotron is the driver machine of BRIF. The working

drawings of the cyclotron magnet were finished and the RF powers and RF transmission line were in the process of ordering. The test bench for central district of the cyclotron was under construction.

2 Isotope separator on line (ISOL)The target-source of ISOL would be a very strong radioactive device after 100 MeV, 200 µA on the

target for a period of time. The special radiation protection measure for the target-source room should be taken and was designed in 2006. Two 20 cm thick ion-lead doorlike movable plates were adopted. It ’s open for target-source routine operation and close for maintenance. The preliminary design of hot cell dedicate to target-source repairing was completed too.

3 Superconducting boosterThe 2 MeV/q superconducting booster was in the stage of construction. The cupper Quater-Wave

Resonators (QWR) and the cryostat were under machining. QWR will be ready for Ni-sputtering and testing in 2007.

4 InjectorThe injector up grading program includes promoting injector platform high voltage from 150 kV to

300 kV and building a new injection system dedicate to AMS with mass resolving of better than 380. The program finished on October in 2006 and now is in test operation. The mass resolution for AMS is better than 480.

Progress on Design and Construction of CYCIAE-100

Technology Division of BRIF

13 Annual Report of China Institute of Atomic Energy 2006

The project of 100 MeV high intensity cyclotron and its beam transport lines, CYCIAE-100, is designed to provide a 75-100 MeV, 200-500 µA proton beam with 7 proton beam lines and 2 neutron beam lines. In 2006, the preliminary design has been accomplished and has been tered into the phase of detailed design and construction. The fabrication of some key facilities and devices has been carried out and the R and D items on some design verifications have been basically finished.

1 Sign of CYCIAE-100In 2006, based on the preliminary design of CYCIAE-100 in 2005, the technology division of BRIF

carried out a more detailed design according to the technique features of the project. The preliminary design of the following 24 sub-systems has been finished: A02 general layout of BRIF; B00 general design of CYCIAE-100; B01 the main magnet; B02 the main coil; B03 RF system; B04 ion source, injection line and central region; B05 dual extraction system; B06 Beam development and beam dump system; B07 vacuum system; B08 cooling system; B09 pneumatic system; B10 beam diagnostics system; B11 mapping and shimming system; B12 hydraulic pressure raising system; B13 transportation of large pieces; B14 computer control system; B15 power supply system; B16 electric system; B17 general assembly and maintenance system; B18 Commissioning; B19 dose monitoring and safety interlock system; B20 magnetic field adjusting system; F01 proton beam lines; F02 neutron beam lines. Besides, we also participated in the engineering design of radiation protection and monitoring system (A04), and the writing of the new research report of feasibility and the safety analysis report. We also took the leading role in the preliminary work on the engineering design of the overall control system of the project (A03). As a consequence, we finally formed a 24-volume file on the engineering design, which are over 950 pages in total. The drawings are in two volumes, over 80 assembly drawings for the sub-systems. The simulated assembly of the main parts of the 100 MeV cyclotron is shown in Fig. 1. Apart from those listed above, we also finished 6 versions of budgetary estimate files and provided them for review.

Based on the engineering design, we started the detail design and construction. Our emphasis is on some key sub-systems such as the main magnet and the RF system, some parts requiring sophisticated technique and long-time research such as beam diagnostics, ion source, injection and central region. The requirement for the on-site installation was also investigated, especially for the main magnet, vacuum leaking detection, magnetic field mapping and shimming. This requirement has been delivered to the construction unit of civil engineering.


Fig. 1 Simulated assembly of main parts of CYCIAE-100

2 Construction of main parts/devicesThe construction of the main parts/devices has been started and it mainly includes the following.1) Carried out work on the magnet design solution, fabrication techniques and manufacturer

selection. We’ve already signed the contract with Industeel, France for the hot rolled carbon steel plates for the main magnet (8 sectors), ultrasonic flaw inspection, and magnetic properties test. According to the contract, the materials for the 8 sectors will arrive at Tianjin Port, China in November, 2007.

2) Finished the construction design of 2 100 kW RF amplifiers and the power transportation system, the tendering and bidding work of the amplifiers and the transportation system, and signed the contract with China Aerospace Science and Industry Corporation for the collaborative fabrication of relevant devices.

3) Accomplished the research of the 10 mA experimental ion source prototype.4) Carried out the research on the digital low level RF control system, as it is shown in Fig. 2. The

phase stability is better than 0.3, and the amplitude stability is better than 2‰. It will be improved further more.

5) Carried out the research on beam double-wire scanner and DCCT.6) Finished the research and technical experiment on the radiation-resistant insulate materials for the

tuning coil for the isochronous field and centering coil attached on the surface of the poles in the central region.

Fig. 2 Digital low level RF control system

3 Progress on experimental verification3.1 Test stand of Central Region Model (CRM Cyclotron) for 100 MeV cyclotron

The experimental magnet of the CRM Cyclotron arrived at the Cyclotron Lab of CIAE on March 24, 2006. The magnetic field mapping and shimming was accomplished at the end of June of the same year. By the end of 2006, all the hardware of CRM Cyclotron had been finished, including 1 set of 4.55 T·m 4-sector magnet of 13 tons, 2 exciting coils, two 70 MHz resonance cavities, 1 set of 10 kW RF power supply, 10 mA H-ion source, injection line and central region, stripping extraction system, vacuum system, power supplies, pneumatic system, water cooling system, control system, dose monitoring and safety interlock, etc. The fabrication had been finished and the installation was done successfully. It marks the accomplishment of the construction on the test stand of the Central Region Model. And we are able to


carry out part of verification work for the 100 MeV cyclotron by this test stand. Through the magnetic field mapping and shimming, we verified the crucial design of raising vertical focusing by consistent hill gap adjusting in the 100 MeV machine. In addition, we investigated a new method for the 1st harmonic shimming. The effectiveness of this method was verified in the magnetic field shimming of CRM and consequently reduced the technical requirement for the top/bottom and return yokes. The test stand and part of relevant devices of CRM Cyclotron are shown in Fig. 3.

Fig. 3 Picture of CRM cyclotron

3.2 Experimental verification on RF resonance cavityCurrently we have finished the design, processing and fabrication of the 1:1 scale cold model cavity,

accomplished the test of the key parameters such as resonance frequency and accelerating voltage. The investigation on the moving and electric contacting material was carried out and part of the welding experiment has been finished for the 1:1 scale RF metal experimental cavity. The first edition of the fabrication drawing for the metal experimental cavity has been accomplished. Through the wide and intense contact with the manufacturers and in-depth discussion on the fabrication consideration, we have selected the adequate manufacturer and signed the mechanical processing contract accordingly.

In general, it is proceeding smoothly on the design, construction and experimental verification of the CYCIAE-100 project. The working emphasis for 2007 will be the construction of the main magnet, the processing of the RF cavity, the construction design and starting the fabrication on the vacuum system.

(Written by ZHANG Tian-jue, LI Zhen-guo, CHU Cheng-jie)

Comparison of Design Alternatives and Selection of Manufacturers for Main Magnet Construction of CYCIAE-100

ZHANG Tian-jue, CHU Cheng-jie, ZHONG Jun-qing, CHEN Rong-fan, LU Yin-long,YANG Jian-jun, SONG Guo-fang, ZHOU Zheng-he, LI Zhen-guo, XING Jian-sheng,

WEI Su-min, YAO Hong-juan, LIN Jun, JING Wei, CAI Hong-ru, YANG Fang, WANG Zhen-hui


The 100 MeV high intensity cyclotron, CYCIAE-100, is an advanced accelerator project in China. The fabrication of the main magnet is the key part of the construction for CYCIAE-100 and the technology division of BRIF is consistently taking a discreet attitude towards the magnet design. From the year 2004 to 2005, we carried out intensive investigation on the deign features of the magnet of cyclotrons in institutions and companies like TRIUMF (Canada), PSI (Switzerland), GANIL (France), Cyclotron Advance (Canada), and so on. We designed five different schemes for the main magnet of CYCIAE-100. The major work in 2006 concerning the magnet is emphasized on the comparison of design alternatives and selection of its manufacturers.

1 Selection of design alternativesIn April, 2006, following the evaluation on the magnet by domestic experts, we chose two

alternatives out of five as the major solutions. In July, while visiting TRIUMF, we brought the preliminary design drawings concerning the two solutions with us and consulted relevant specialists on cyclotrons and magnet design at TRIUMF, including Drs. Mike Craddock, Ewart Blackmore, George Mackenzie, Thomas Kuo, Alan Otter, George Clark. Besides, the in-depth discussion was also carried out with experts from unit other than TRIUMF such as Dr. Bill Gyles. In November, we convened a domestic expert panel discussion on the magnet design and its fabrication techniques. In the discussion, detailed reports were given respectively by designers from CIAE and Capital Aerospace Machinery Corporation. The experts participated are from a wide range of disciplines, including magnet design, metallurgy, mechanical processing, installation, et al. In the end, the structural design of the magnet was determined and the following are the highlights:

1) Compact integral structure, straight edge sector, consistent changing hill gap of sector magnet;2) The core parts, such as the sectors, take use of forged pure iron. The top and bottom yokes and

return yokes adopt GB 8# Steel in the light of China standard and casting piece can be considered. The uniformity requirement for carbon content should be within the range of 0.02%-0.06%, and the internal imperfection should be less than 8 mm for the casting steel;

3) The hill gap tolerance should be 0.1 mm after installation and assembling. After the finishing machining and installation of the four sectors, the combined machining of the pole surface, the contacting surface between the top/bottom and return yokes should be done to eliminate tolerance accumulation;

4) In order to reduce hill gap deformation caused by atmospheric pressure, the weight of the magnet and electromagnetic force, we decide to adopt top and bottom yokes with different height along the radius.

2 Risk analysis and protective measuresIn the phase of engineering design and alternative selection mentioned above, many crucial

technological problems have been researched and solved. There are no subversive technical risks in this process. However, since the project has entered into the phase of detailed design and construction, we should get fully aware of various risks and challenges that might rise in the process, and it is necessary to take initiatives to find corresponding measures to the problem.

1) To cast the top and bottom yokes as an integral part, there might be risks in the aspect that the carbon content segregation and inner shrinkage porosity can not meet the technical requirement. In the respect of shrinkage porosity, currently three of the domestic manufacturers conclude that there is difficulty to control the inner imperfection within 1-3 mm, while it is possible to reach the requirement for 3-5 mm according to their quality control plan. Through careful design and analysis, our technical requirement for the manufacturers is 8-12 mm, and if necessary, we can consider weld repair based on


the alternative agreed by both sides. For the carbon content segregation, it is obvious that 0.01% is hard to achieve. Currently the responding from two manufacturers is 0.02%-0.05%, while our technical requirement is that for similar radius, it is 0.02%-0.05%, and for different radius, the segregation should meet the requirement of GB 8# Steel. Even if both of the two aspects above can not be controlled and break the bottom line for the technical requirement, on the premise that the forging materials (poles, et al.) meet our requirement, the influence of shrinkage porosity and carbon segregation in the top/bottom and return yokes to the magnetic field in the hill gap will be mainly average effects. According to our design, the tuning coils on the return yokes outside of the vacuum chamber can adjust the magnetic field of each sector in average.

2) The machining tolerance will directly influence the isochronous field and harmonious field. The main tolerance such as the profile tolerance of the hill gap is 0.1 mm. In case this requirement can not be reached, for instance, it is increased to 0.2 mm in practice, according to the real tolerance situation, we can take advantage of mending, magnetic mapping and shimming to meet the requirement of the cyclotron field.

3) The risk exists that the axial focusing force might be not strong enough by this straight edge sector. The main factors that influence the axial focusing force for CYCIAE-100 are BH curve shift and the pole angle positive tolerance. If the BH curve does not meet the requirement, we can solve it by increasing the excitation current of the main magnet and reducing the angular width of the shimming bars. The present experimental research demonstrates that it is feasible to reduce the magnetic premeability by 5%. If it is the positive tolerance of the pole angle, it can be solved by refabricating the side of the sector or reducing the angular width of the shimming bars.

Therefore, though there are several risks mentioned above, they can be solved by proper measures and will not lead in significant change for magnet construction in the process of building the cyclotron.

3 Manufacturer selectionThe plan on the magnet construction of CYCIAE-100 adopts design, installation, assembling, field

measuring, and shimming by CIAE and meanwhile relies on the support from heavy machining factories home and abroad to accomplish the metallurgical and machining work.

The upper half The lower half

Fig. 1 Main magnet of CYCIAE-100

After several negotiations with the possible collaborative manufacturers in and outside of China, including Industeel in France, China First Heavy Industries (CFHI), China National Erzhong Group Co. (Former The Second Heavy Machinery Plant, hereinafter abbreviated China Erzhong), and Capital Aerospace Machinery Corporation, it has been determined that Industeel will take the work of providing


the materials for the sectors, and the contract has been signed. It is estimated that the materials will arrive at Tianjin Port, China in November, 2007. The providers for the top/bottom and return yokes will be selected from two domestic manufacturers soon. The final negotiation with them is in progress. We plan to buy directly from Taiyuan Iron & Steel (Group) Co., Ltd. (TISCO) the pure iron materials for other parts of smaller dimension like the central plug and shimming bar.

The main magnet of CYCIAE-100 is shown in Fig. 1.

100 MeV Cyclotron LLRF Control Design and Details for RF Circuit

YIN Zhi-guo, XIA Le, HOU Shi-gang

The research on the RF control of the 100 MeV cyclotron has entered into detail design phase since 2006. In order to minimize the design risk, two VXI C size homemade control boards were produced and tested for certification purpose, together with related software and firmware. Although with the absence of cavity, in the test setup, the controller gain was limited to around 1.8, and the system functions well as expected. The most important thing is that, through this test, we gain more confidence about the substitution of digital controller to the traditional one. The simplified LLRF system block diagram is shown in Fig. 1. Related detail information for RF part and PI regulator part can be found in this paper and “Development of VXI-bus based DSP board”.


Fig. 1 100 MeV LLRF test system block diagram

1 Detailed RF circuit design1.1 Amplitude and phase demodulation.

Form the perspective of control, LLRF system performance and accuracy are strongly depending on feedback signal demodulation.

There are many choices for RF amplitude detection, but traditional signal diode option[1] was rejected due to its nonlinearity. After experimental comparison, we decide to take the change from compensated double diode detection, which had been adopted for several LLRF systems[2] to the double balanced mixer based one[3]. For D.B.M. amplitude detection the difficulty is to re-generate highly stabilized (both in amplitude and phase) local RF signal as a reference. A high performance [4] limit amplifier AD8306 form Analog Device was utilized to achieve this goal.

There are also alternatives for Phase/Frequency detection, an advantage taken from the maturity of PLL technology. For the next step, smoothly moving to “Cavity self-excited phase lock mode” [5], two different types of phase detectors were adopted for the design, one is AD9901 form Analog Device, and the other is MCH12140D from ON Semiconductor. 1.2 Vector modulator

Vector modulator is a device in which the input signal is split into a pair of orthogonal components normally designated as I (in phase) and Q (in quadrature phase). The amplitude of each component is independently controlled through a couple of linear attenuators; then the output signal is obtained by summing in-phase the I and Q components. Four quadrant analog multipliers (AD834) from Analog


Device were adopted as core components, for its exceptional high bandwidth (up to 500 MHz) with a good linearity and constant phase response[5].1.3 Bandwidth and other issues

System bandwidth, noise level and time delay are always conflict goals for electronic systems which is especially obvious in LLRF system. The regulator time delay will ultimately limit the controller stable gain, which will in turn limit the accuracy of the system. To have a trade off between these goals, 100 kHz has been selected as the system bandwidth for this test, which is sufficient when finally run with high Q cavity present. From this point of view, all Anti-aliasing and anti-imaging filters in this design have been simplified to 2 stages. Some of them were even cancelled due to the fact that there were enough poles contributed by OPAMP on the signal path.

2 Test resultThe preliminary desktop test includes two phases: open loop test and close loop test. Both tests show

these two boards function as designed. For the purpose of system stabilization, high controller gain test was not performed this time due to the absence of RF Cavity or equivalent circuit. In spite of that, during the 1.5 h test, the controlled RF signal can be stabilized within 1.2 mV amplitude variation and 0.2 degree phase variation.

References:[1] ZHAO Zhenglu, et al. CYC30 RF system, Internal Report.

[2] CARUSO A, et al. The upgraded control of the LNS Superconducting Cyclotron Radio-Frequency system. LNS

Internal Report.

[3] LAVERTY M, FONG K, FANG S. TRIUMF ISAC II RF control system design and testing. EPAC04.

[4] ZHENG Qiwen. High performance limiting-logarithmic amplifier, TRI-DN-02-05. 2002.

[5] BASSATO G., PONCHIA R. Design of a low cost vector modulator for signal processing applications in the frequency

range 1 to 200 MHz. LNL Annual Report 2002.

Development of VXI-Bus Based DSP Board

HOU Shi-gang, YIN Zhi-guo, XIA Le, JI Bin

As it has entered into detail design phase for the RF control of the 100 MeV cyclotron, to minimize the design risk, two VXI C size homemade control boards were produced and tested for certification purpose, together with related software and firmware. One of them is VXI-bus based DSP board. Inside this board, two Freescale DSP-56303 digital signal processors were utilized to archive 3 channel PID regulation. Preliminary test shows this module, together with its VXI bus interface, will meet the needs of LLRF control for cyclotrons, both from perspectives of flexibility and optimization.

The key feature of DSP board is:1) VXI Bus based design, which gives the system more flexibility from the point of computer

control.2) Digital signal process based design, which gives the core part more immunity to temperature drift,

noise, et al., compared with analog one.


3) Easy to optimize loop parameters. The communication via HI08 port of DSP56303 makes this design supported with on-line adjustment. Dynamic change of “P” and “I” parameter is possible without any modification to hardware circuits.

4) Other operation modes like adjustable Pulse mode, Ramping RF power, Open/Close loop functions, re-calculate PI parameter et al. can help for running RF system as well as RF conditioning process.

The draw back of digital PI controller is, sometimes in loop regulation, the error signal is so small that it is very difficult to be quantized. To solve this problem, a 18 bits analog to digital converter from Analog Device were select in this design, which can theoretically provide 102 db dynamic range. For the opposite procedure, driver ability is more emphasized. AD7840 were used as digital to analog converter for its driver ability and its good signal to noise ratio. As 100 kHz system bandwidth has been adopted for the test, by Nyquist-Shannon sampling theorem and from the perspective of practice, 500 kHz sample rate was adopted to ensure the system bandwidth. Architecture of this digital regulator is shown in Fig 1. To get the maximum system performance four layer PCB design is adopted, which also helps to reduce noise coupling and impedance control for important signals. The picture of PC Board is shown in Fig. 2.

Fig. 1 Architecture of digital regulator

System test has been performed together with RF board, and the result shows that both board function as designed. Loop regulation bandwidth were identified by using manually sweeping sinusoid wave method, through which, pole and zero can be easily observed and changed online.

Fig. 2 LLRF DSP board

Development of PROFIBUS Based Control System

YIN Zhi-guo, HOU Shi-gang, LI Zhen-guo, ZOU Jian


We designed control system for 100 MeV cyclotron, which include device layer, front end computer (FEC) layer and operator layer. This architecture is accord with standard control model of accelerator.

There are almost 1 000 signals of all cyclotron subsystem, including 250 analog input signals, 250 analog outputs and digital I/Os. According to the operating experience of 30 MeV cyclotron, we know that the device of cyclotron should be controlled and supervised frequently, but the communication data are few. PROFIBUS-DP was selected as communication protocol between controllers, field devices and FECs. PROFIBUS-DP is a member of PROFIBUS protocol family, which aims at the high-speed data exchanging between separated devices. In field device layer SIEMENS PLC series are adopted as controllers. The architecture of device layer is shown as Fig. 1.

Fig. 1 Device layer based on PROFIBUS

There are four controllers in device layer which have different modules. The CPU are S7-400 series, these controllers communicate with other devices via PROFIBUS. At runtime, one of these controller will act as a master station on PROFIBUS network, the others work as slave station, the signals controlled by salve station will be mapped to master station.

Each controller include one or more CP342-5/CP442-5 PROFIBUS module, this module achieve the communication between FECs and PLCs. A CP5613 card was installed on each FEC, so FECs can work as master station on network. We implement a 2-node PROFIBUS on centre region prototype control system, when the communication rate at about 20 Hz, the length of data frame is 240 byte, the load of CPU no more than 3%,this performance will be enhanced in the future if needed. It will be save time and expense on connecting devices, field bus can void the interference between devices. It is suitable for cyclotron control system.

Calculation of Transverse Acceptance of CYCIAE-100

YAO Hong-juan, ZHANG Tian-jue, LU Yin-long

The acceptance of central region should be considered after the central ray is finished. This can check whether the central region design is reasonable or not, because in practice it is the bunched beam instead of the single particle that is accelerated; on the other hand the results of acceptance calculation


can be provided as fitting conditions to injection design, and making the match calculation of injection line easier. The central region structure of CYCIAE-100 and trajectories of reference particle with 40° phase width is shown in Fig. 1 (left), the right one is phase history with energy.

Fig. 1 Trajectories of reference particle with 40° phase width and phase history

▲——185(15)；◆——165(－4)；■——145(－19)

Based on the results above, we selected a number of particles (with different r and Pr)，and tracked forwards starting from injection point with 40 keV to 1MeV. During the acceleration process, we removed the particles hit on electrodes, then at 1 MeV we checked whether the particles unlost were inside the static ellipse or not using three different normalized emittances of 1π mm·mrad, 2π m·mrad, 3π mm·mrad. From compute code output we can get the phase coordinates of particles at 1 MeV, and then recorded the particles which were inside the static ellipses; see Fig. 2, for the central phase.

Fig. 2 Particles were accepted at 1 MeV for central phase

Real line——en=1π mm·mrad；Broken line——en=2π mm·mrad；Dotted line——en=3π mm·mrad；◇——Inside 3π mm·mrad；▽——Inside 2π mm·mrad；◆——Inside 1π mm·mrad

Before tracking we marked all particles, so at injection point the coordinates of particles which were


accepted at 1 MeV can be obtained. Calculations were done for central phase and plus and minus 20° from the central phase. We recorded particles accepted for different phases and fit the overlap using the ellipse, and then the radial acceptance was obtained. The same calculation was done for vertical phase space (z-Pz). We selected the normalized emittance of 2 πmm·mrad for CYCIAE-100, which agrees with the requirements of extraction design. The results are shown in Fig. 3, radial acceptance is about 0.4π mm·mrad, and the vertical one is about 0.3π mm·mrad

Fig. 3 Radial and vertical acceptance results (at 1 MeV by normalized 2π mm·mrad)

◇——165（－4）；□——145（－19）；△——185（15）

Stripping Extraction Simulation With COMA for CYCIAE-100

AN Shi-zhong, ZHANG Tian-jue, GUAN Feng-ping, ZHONG Jun-qing, JIA Xian-lu,WEI Su-min, YAO Hong-juan, BI Yuan-jie, YANG Jian-jun

With multi-particle tracking code COMA, the process of stripping extraction for CYCIAE-100 is simulated and the position of stripping foil, which is got from the code CYCTR, is tested by COMA. Beam dynamics are studied in detail and the beam parameters after stripping are analyzed.

The H-beam is injected from the symmetry center of valley with azimuth θ=0°, and the beam will be tracked along the inserting direction of stripping probe with azimuth θ=61.02°. The initial beam parameters: E0=1.49 MeV, R=23.1 cm, phase extension in RF with Δφ=±20°, normalized transverse emittance is 0.01 cm2. The input phase distributions are uniform in both transverse and longitudinal directions and 4 920 macro particles are used. For extraction energy of 70 MeV, the stripping foil is at the radial position of 1.60 m.

1 Process of stripping extractionFor the initial bunch with bunch length of Δφ=±20°, the accelerating process is simulated with

COMA. The average energy of the beam is going up to 91.7 MeV and the average radius of beam is going up to 1.795 m after 281 turns. For the beam with average energy of 70 MeV, the average extraction radius


is 1.6 m. This keeps the agreement with the results got from the code CYCTR. From the simulation results, the average extraction energy of the beam is 70 MeV when the stripping

foil is fixed at the radius of 1.6 m. The protons begin to be extracted by charge exchanging when the H-beams go through the stripping foil after the 208th turn and the whole H-particles in one bunch will be extracted after 80 turns. It means that it needs about 80 turns for one bunch to be extracted completely, so many extraction turns are due to the large energy deviation in one bunch produced by large initial phase width. Figure l shows the beam phase space distribution and the pink part are the extracted particles after stripping foil. Figure 2 shows the extracted phase space distribution after stripping foil with multi-turn extraction. The elliptical phase space boundary, the yellow line, is got from 5×RMS realistic emittance. The extracted trajectory calculation after stripping foil will be based on this extraction distribution. From Fig. 2, all of the extracted particles are inside the elliptical phase space boundary.

From Fig. 2, the beam envelope after stripping extraction in transverse is about less than 1 cm and the extracted beam energy spread is about±0.5%. The normalized emittance is roughly unchanged and the bunch length is about 55° for the extracted beam. From extracted beam distribution, the size of stripping foil needs at least 1 cm in both radial and vertical directions.

Figure 3 shows the multi-turn extraction process. The extracted beam is about 1% of one bunch after the 208th turn and it is up to 100% after 80 turns. Figure 4 shows the extracted beam profile. It is Gaussian like distribution and the particles between the radius of 160 cm and 160.4 cm are about 96% of one bunch.

2 Extracted beam parameters From the simulation results with COMA, the extracted beam parameters after stripping foil:the

average extraction energy is 70.54 MeV for the average extracted radius of 160.23 cm; the average RF phase is 34.9°. The bunch center is at x=－0.05 cm, Px=0.1 cm or 0.6 mrad and z=－0.000 2 cm, Pz= －0.003 cm or –0.017 mrad. The normalized transverse emittance is roughly unchanged and the beam RMS envelope is xrms=0.12 cm, zrms=0.14 cm. the maximal phase width is 55° and the maximal energy spread is 1%(±0.5%).

Fig. 1 Phase space distribution after the 208th turn


Fig. 2 Extracted phase space distribution with multi-turn stripping extraction

Fig. 3 Extracted particle numbers with multi-turn extraction process

Fig. 4 Extracted beam profile

Accelerating Trajectory Simulation With COMA for CYCIAE-100

AN Shi-zhong, ZHANG Tian-jue, GUAN Feng-ping, ZHONG Jun-qing, JIA Xian-lu,


WEI Su-min, YAO Hong-juan, BI Yuan-jie., YANG Jian-jun

With multi-particle tracking code COMA, the process of accelerating for CYCIAE-100 is simulated and the beam dynamics such as the energy spread, phase slip, emittance, envelope etc. of the beam with the acceleration are studied in detail.

1 Input conditionThe H-beam is injected from the symmetry center of valley with azimuth θ=0°, and the beam will be

tracked along the same center of valley. The initial beam parameters: E0=1.49 MeV, R=23.1 cm, phase extension in RF with Δφ=±20°, with the phase center of φCP=－13.5°, normalized transverse emittance is 0.01 cm2. The input phase distributions are uniform in both transverse and longitudinal directions with the initial zero energy spread and 4 920 macro particles are used. The tracking turn is 275 in total.

2 Results in longitudinal direction For the initial bunch with energy of 1.49 MeV, accelerating process is simulated with COMA. The

average energy of the beam is going up to 82 MeV and the average radius of beam is going up to 162 cm after 275 turns. The phase center of the bunch is at φCP=7.5° and the phase slip is up to 20° after 275 turns, but the total integral phase slip is close to zero. The phase width of the bunch is almost unchanged and the maximal energy spread is 7 MeV after 275 turns. So large energy spread is mainly due to the large initial phase width and the phase slip of 20°. The longitudinal results are shown in Figs. 1-3.

Fig. 1 Average energy vs turns and radius

Fig. 2 Phase slip during the accelerating


Fig. 3 Longitudinal phase space distribution

3 Transverse resultsThe normalized emittance is unchanged in z direction during the acceleration, but it is a little bit

reduced in radial direction due to the adding the accelerating voltage with the larger radius. The envelope and the beam center in x and z directions are almost unchanged, which means the beam is matched very well. The realistic emittance is reducing along the radius and the transverse phase space is reducing with the adding energy of the beam. Due to the periodic symmetric focusing feature of cyclotrons the transverse phase ellipse in the center of valley is upright. The transverse results are shown in Figs. 4-6.

Fig. 4 Transverse phase space is reducing with the radius

Fig. 5 Normalized transverse rms emittance


Fig. 6 Transverse beam rms envelope

PIC Code for Beam Dynamics Simulation and Its Applications on Injection Line of CYCIAE-100

YANG Jian-jun, ZHANG Tian-jue, AN Shi-zhong, JIA Xian-lu, QIANG Ji1, LIN Yu-zheng2

(1 Lawrence Berkeley National Laboratory; 2 Department of Engineering Physics, Tsinghua University)

The space-charge effects can impose significant influence on the high intensity beam transport process. Along with the development of computing physics and computer science, PIC (Particle-In-Cell) methods are playing an important role in the accurately numerical simulation of charged particle beam’s behavior in accelerator and beam transport line.

In order to carry out an in-depth study on the negative hydrogen DC beam transport on the axial injection line of CYCIAE-100, a high intensity cyclotron is under design at CIAE. We developed a new two-dimensional PIC code CYCPIC2D, applied it to simulate high intensity DC beam transport on the injection line, and compared with the results given by ORBIT and TRACE-3D.

1 Numerical methodsThe general form of potential solution of two-dimension poisson equation of electric field can be

expressed as:

(1)

Where (x, y) and (x, y) are the coordinates of target point and source point respectively, G(x, x, y, y) is Green function, and (x, y) is the charge density.

In our code, the computational domain including all particles is divided into a discrete grid space and the particle charge is deposited onto the grid nodes using a CIC (Cloud-In-Cell) scheme to obtain the charge density of the grid space. Then the potential at the grid nodes can be expressed as:

(2)

Where Δx and Δy are the mesh sizes along x and y directions respectively, and Nx and Ny are the number of nodes along x and y directions respectively.

The space-charge potential on the nodes can be calculated by solving Eq. (2) using a FFT (Fast Fourier Transform) based algorithm given by Hockney[1]. Then the electric fields on the grid are


interpolated to obtain the space-charge force on each macro-particle.To integrate movement equation of charged particle, symplectic split-operator methods are adopted

to separate external applied field and space-charge effects [2]. In the each step, the beam distribution of particles is transported through the first half of the step in the applied field firstly, then the space-charge effect is added through full steps, and lastly, the new beam distribution of particles is transported through the remaining half of the step in the applied field. In our code the longitudinal coordinate is used as the independent variable.

2 Object-Oriented programmingFortran is a widely used high effective programming language. We adopted Fortran 95 and Object-

Oriented Designing methods in our code. After the specific physic problem was decomposed and classified, different modules (as Class in C++ language) were set up by abstracting different aspects of the problem. Each module consisted of derived data types (type) and operations on it (subroutine and function). The modules can be classified into three types from bottom to top: functional modules, physical modules and applicative modules. At debugging phase, each module was tested separately and then assembled into the whole program to ensure validity of the code.

3 Simulation of the injection line of CYCIAE-100The DC beam extracted from ion source is injected into the middle plane along the axial coordinate

of the machine. The arrangement of transport elements is O-BS-O-D-O-F-O (BS: magnetic solenoid, D: defocus quadrupole, F: focus quadrupole and O: shift tube). The parameters of each element were determined by TRACE 3-D code.

As the first application case, we simulate the beam transport from the exit of the ion source to the entrance of spiral inflector under different neutralized rates. The result is also verified by being compared with that from ORBIT and TRACE-3D. The RMS beam envelops given by the three codes are shown in Fig. 1.

As shown in Fig. 1a, envelops from the three codes agree well with each other at highly neutralized rates, which means their precisions are basically identical in low intensity conditions. CYCPIC2D and ORBIT can still give almost the same envelops under highly neutralized rates, while the result of TRACE 3-D becomes divergent, as shown in Fig. 1b and c. When the beam intensity improves to 20 mA, the divergency of TRACE 3-D is particularly evident, as shown in Fig. 1d. This phenomenon is caused by the error introduced by the truncation of space-charge force to linear terms in TRACE-3D, which gets bigger and bigger along with the increase of beam intensity. In addition, the maximal value of RMS envelops for 8 mA beam with neutralized rates of 90%， 50% and 0% are 1.57, 2.94 and 4.35 cm, respectively. Whereas, due to the space limit in the injection region of the machine, the designed wall radius of the tube is only 2.5 cm. Therefore, in order to achieve a reasonable transport efficiency, it is critical to keep the neutralization high enough in the tube of injection line.


Fig. 1 RMS envelop given by CYPIC2D, ORBIT and TRACE 3-D

a, b, c——8 mA beam, 90%，50% and 0% neutralized rates; d——20 mA beam，no neutralization

References:[1] HOCKNEY R W, EASTWOOD J W. Computer simulation using particles. New York: Hilger, 1988.

[2] FOREST E, et al. Phys Lett, 1991, A158, 99.

Study and Practice of Shimming Method to First Harmonic in Compact Cyclotron

ZHANG Tian-jue, ZHONG Jun-qing, WANG Chuan, LU Yin-long,WANG Zhen-hui, YANG Jian-jun, LIN Jun

Because of the intrinsic asymmetry of the magnet，imperfection of the machining and assembly of the magnet, there is imperfection distribution of field in the median plane of compact cyclotron, and this kind of imperfection will include the various harmonics by Fourier analysis. Due to the existence of the first harmonic field, the envelop of particle will be distorted and oscillated, and meanwhile it will bring the offset of the center of beam trajectory. All these lead to the fact that the first harmonic will impose significant affect on the quality of the accelerated beam. As a consequence, not only the isochronous but also 1st harmonic is shimmed in the cyclotron. However, considering the factors inducing 1st harmonic and the function limit of shimming bars to the first harmonic, the first harmonic will not be eliminated by shimming bar machining.


1 The first harmonic of the magnetic field in the Central Region Model (CRM)The CRM of 100 MeV cyclotron is used to carry out experiments on the high intensity beam. The

intensity of the beam will be limited due to beam losses resulted from the big 1 st harmonic field. Consequently, the amplitude of 1st harmonic at small radius and big radius should be less than 15 Gauss, while at others radius the amplitude is required to be less than 5 Gauss in the CRM.

Fig. 1 Amplitude of the first harmonic in CRM cyclotron

It took several times to measure and shim the isochronous field of the CRM, and after the isochronous field became basically meet the requirement as designed to match the accelerated particle, we got the amplitude of the first harmonic in various radiuses by Fourier analysis, and the result was shown in Fig. 1 of solid line. As can be seen in the Fig. 1, the amplitude at small radius and big radius exceeded the requirement of the accelerated particle, and the amplitudes at a number of others radius also exceeded the requirement.

2 Study of the shimming algorithm to the first harmonic in cyclotronThe first harmonic field exceeded the requirement of the accelerated particle after several times of

isochronous shimming in the CRM, so we should develop an effective method to reduce the first harmonic field. To remove the first harmonic, we shimmed the two shimming bars located closest to the first harmonic. We found the following reasonable prediction from the simulation result of magnetic field and experience of shimming to isochronous of CYCIAE-30.

Where, B(r) is increment of the average field in radius when shimming bar increases 1° in azimuth.In radius, when there is the first harmonic with an amplitude of H1 and a phase of θ in hill of field,

four shimming bars will be shimmed, one couple of them are located in pole edge shown in Fig. 2 and the other couple are located in opposite pole edge. The increments of shimming bars are as follows.

Where, K is shimming factor, with the range from 0 to 1. The increments of opposite shimming bars are －dα1 and －dα2 respectively.


In radius, when there is the first harmonic with an amplitude of H1 and a phase of θ in valley of field, four shimming bars will be shimmed, which are located in 4 pole edges respectively, one couple of them are shown in Fig. 3 and the other couple are located in opposite pole edge. The increments of shimming bars are as follows.

The increments of opposite shimming bars are －dα1 and －dα2 respectively.

Fig. 2 Sketch map of 1st harmonic in hill

3 Result of the shimming to the first harmonic fieldWe obtained four shimming curves of four bars by selecting different shimming factors in different

radius. The increments of two shimming bars were plus in numerical value, which means some iron should be glued to shimming bars, but this method was impracticable in the CRM. Considering that the increment of the shimming to the first harmonic is small, we shimmed the isochronous field and the first harmonic at the same time after several times of shimming the isochronous. In the end, the plus increment did not happen. The result of shimming to the first harmonic is shown in Fig. 1 in broken line.

As shown in Fig. 1, we know that this method of reducing 1st harmonic by shimming bars is effective, and it makes it possible to change the phase of 1st harmonic.


Fig. 3 Sketch map of 1st harmonic in valley

Method on Adjustment of Isochronous Field Change Causedby Temperature Increase for CYCIAE-100 and

Experimental Study on Coil Insulation

WEI Su-min, ZHANG Tian-jue, CHU Cheng-jie, YANG Jian-jun, Ge Tao, ZHOU Ke, LU Yin-long,ZHONG Jun-qing, YAO Hong-juan, JIA Xian-lu, SONG Guo-fang

In the compact cyclotron with small magnetic pole gap, it places high requirement for the isochronous filed, as is the case with the 100 MeV high intensity cyclotron under construction at CIAE. As the gap between the sectors is small, magnetic field in the gap is sensitive to the magnetic pole deformation caused by factors such as temperature increase and electromagnetic force while the machine is in operation. In this case, the isochronous filed should be maintained by compensating the field divergence caused by pole deformation on line to guarantee the normal accelerating of the particles. After wide surveys and detailed calculations, meanwhile with an insightful comparison of the frequently used adjusting solutions on the isochronous filed around the world, the adjusting capability, adjusting precision, technical feasibility and the building cost of control system of “trim coil” scheme and “trim rod” scheme, we finally chose the former by installing 4 sets of isochronous field adjusting coils in a pair of opposite valleys and 8 sets of centering coils on the magnetic pole surface in the central region. The models of those coils are shown in Fig.1.There are 8 power supplies for the isochronous field adjusting, and 9 power supplies provide power for the centering in the central region. Through changing the different combination of the current of coils, different magnetic field can be obtained to compensate the corresponding field divergence in various situations.

During the operation time, the heat from other parts of machine, such as RF cavity and the main coil, will be partly transmitted to the magnet and cause magnet deformation and further influence on the main magnetic field. Figure 2 shows field divergence from isochronous field caused by temperature increase of the magnet by 25 ℃ after modest compensation by trim coil, from which it can be seen that the maximum field change is about 25 Gauss in this case. After changing the current loaded on the trim coils, we can adjust the isochronous field back to the value close to the theoretical calculation. In Fig. 2, the horizontal


coordinate is the radial position, and the vertical coordinate is the field divergence from isochronous field. After the adjustment the value difference at radius 20-190 cm is less than 1.5 Gauss.

Fig. 1 Isochronous field adjusting coils and centering coil

Fig. 2 Field divergence from isochronous field by temperature increase of the magnet by 25 ℃

As a result, in principle it is feasible to use such kind of coils to adjust the isochronous field on line. However, since this kind of trim coils for the isochronous field will be placed in the high vacuum environment, it is difficult to solve the problems on techniques of insulation. As the space of the valley is limited, it will be not easy to install the trim coils. During the operation of the machine, since the temperature and force change are influenced by various factors, the case is rather complicated. In view of these, the adjustment to the isochronous field should be done on line, and accordingly the calculation of the dynamic on-line adjustment is also complicated.

In order to overcome the difficulties in fabrication and installation, based on the physical design, we considered to realize the coil insulation by using alumina films. In this method, we adopt the copper conductor with inner water cooling and aluminum is used as the support which is coated with the alumina on the surface. Such kind of design can guarantee the insulation and meanwhile has relatively minor influence to the high vacuum. To prove the property of insulation, we carried out experimental verifications on the alumina film at different thickness. Figure 3 is the model used in the experiment, and the hole is prepared to fix the screw during installation. Due to the difference in the thickness of the films, the color in the model is slightly different from each other. The result shows that this kind of oxidation technique can make the resistance on the surface of the aluminum block reach above 3 MΩ, which meets the requirement for insulation.

To better solve the problem on the synchronous filed change caused by unpredictable factors in the process of on-line adjustment, we developed an on-line adjustment code for the isochronous field on the basis of results given by theoretical calculation. After verification, it shows that the result has only slight


difference with that given by the magnetic field simulation code.Further construction design will be carried out based on this design.

Fig. 3 Model on coil insulation by using alumina films

Reference:[1] ZHANG Tianjue, WEI Sumin, et al. Preliminary design on the magnetic field adjusting system for the 100 MeV high

intensity cyclotron, Inner Design Report, BRIF-C-B20-01-SM. 2006.

Design, Machining and Installation of Injection Line of CRM for CYCIAE-100

YAO Hong-juan, LU Yin-long, ZHANG Tian-jue, LI Zhen-guo, WU Long-cheng, PAN Gao-feng, GUAN Feng-ping, WEI Su-min, GE Tao, WANG Zhen-hui

For the central region model (CRM) of CYCIAE-100, the 100 MeV high intensity cyclotron under design at CIAE, the injection energy is 30 keV; the transverse focusing elements used are one solenoid and two quadrupoles. There are two considerations about axial injection line for CRM, one case with bunchers and the other without bunchers. The optic layout is shown in Fig. 1.

1 Simulation results CW beam injection without buncherThe initial parameters we selected are r=2 mm, r’=16 mrad, I=8 mA, neutralization is 99%. The axial

component of magnetic field is non-uniform through the inflector. TRANSOPTR can calculate spiral inflectors in uniform field, so we must get the transfer matrix from CASINO with real magnetic field, and then input the matrix to optic program to design injection line including the inflector.

The total length is about 2 m; the first drift is 40 cm. The effective lengths of solenoid and doublet are Les=32 cm and Leq=12 cm, the aperture diameter of Q is 5cm, the maximum pole field of two quadrupoles is less than 600 Gs, the maximum of solenoid field is 2 000 Gs. The beam results are shown in Table 1.

2 Simulation results bunched beam injection with buncherThe initial parameters we selected are r=2 mm, r’=16 mrad, I=1 mA, neutralization is 0. The RF

frequency is about 70 MHz; the voltage of buncher is 400 V; the efficiency is 31%. The beam results are shown in Table 2.


Fig. 1 Sketch of injection lines of CRM

Table 1 Beam size at inflector

PositionBeam spot

x/mm x’/mrad y/mm y’/mrad

The entrance of inflector 2.210 126.4 2.107 63.289

The exit of inflector 3.1 82.153 2.09 49.886

Table 2 Beam size at inflector

PositionBeam spot

x/mm x’/mrad y/mm y’/mrad

The entrance of inflector 2.223 124.263 2.17 66.98

The exit of inflector 2.83 81.04 2.04 53.37


3 Machining and installationThe injection system of CRM for CYCIAE-100 consists of injection line, power devices, vacuum

devices, beam diagnostic devices, water-cooled, pneumatic system and so on. The power devices and vacuum devices were bought, and were installed and debugged in our lab by ourselves. The pipes and part of elements (including yokes of solenoid and quadrupoles) were machined in the experimental factory at CIAE. The coil winding was done in our lab. Besides focusing elements, there are also some beam diagnostic elements, e.g. Faraday Cup, double wire scanner and so on. Some were bought directly and some were designed by beam diagnostic people in our lab. All were measured and met the design demand. At present, the installation of ion source and injection line has been finished, and vacuum test has been accomplished too.

Design, Machining and Installation of Inflector and Central Region of CRM for CYCIAE-100

YAO Hong-juan, LU Yin-long, ZHANG Tian-jue, LI Zhen-guo, JIA Xian-lu, GUAN Feng-ping, WANG Zhen-hui, LIN Jun

The operational principle of spiral inflector is that electric force remains perpendicular to the central trajectory. Because of the magnetic filed in central region, the trajectory through inflector will become a spiral line, and the shape of electrodes is also spiral, so the machining must be done by using numerical control machine tool with the number of its axles no less than 4.

The primary parameters of inflector of CRM are shown in Table 1; the orbits of reference particle and four edges of electrodes are shown in Fig. 1.

Table 1 Primary parameters of inflector of CRMInjection

energy/ keV

Height of

inflector/ mm

Electric

radius/ mmTilt: k’

Gap of

electrodes/ mm

Width of

electrodes /mm

Electric field/

(kV·mm－1)

Rotation at

entrance/°

30 29 33 0 8 16 1.76 260

Based on the central ray, four edges of electrodes and the height of inflector, we calculated the data of electrodes’ surface, then using computational code to convert the data of surfaces to those of milling cutter’s path, that is, the rotation of work piece and vertical and transverse shift of milling cutter. The machining of electrode’s surface can be finished by inputting the data to a numerical control machine tool. In order to make sure of the corresponding relation between the two electrodes, we adopted the same path of milling cutter to do the machining.

The design of central region is so complex that many aspects need to be considered, including magnetic map, electric map, initial point selection and so on. The RF frequency of CRM is 70.493 6 MHz, isochronous field is 11.572 35 kGs, and Dee voltage is 50 kV. There are four accelerated gaps. The design of electrodes in central region is limited by the structure of inflector and shimming bars of the main magnet. AUTOCAD is used to design the electrodes, the origin is the center of the cyclotron, the mesh size in small field is 0.03 cm×0.03 cm×0.13 cm, and in large field is 0.05 cm×0.05 cm×0.13 cm.

The initial parameters for particle tracking are as following: x=2.798 48, y=－ 1.958 30, z=0.0,


Pz=0.0, ALPHA=65.553, TAU=130-170, Ek=30 keV. The results are shown in Fig. 2.

Fig. 1 Four edges of inflector and central trajectory of reference particle

Fig. 2 Trajectories of reference particle with 40° phase width

From Fig. 2 we can see the electrodes of central region are very complex and irregular, and the surface finish quality, dimension and position accuracy are required to be strict to requirement. In order to achieve that, the machining must be done using numerical control tools. The electrodes of central region were finished by using 3-axle machine tools. With the limitation of parallelism less than 0.01 mm between the two flanges, the flat surfaces of flanges were finished using abrasive machining, assuring the precision of single part and position accuracy during installation. Since the position of Dee tip and the size of gaps are very important, a special device was made to guarantee the installation precision, and the error of gaps was controlled within ±0.05 mm. The installation of inflector and central region had been finished, as is shown in Fig. 3.


Fig. 3 Picture of inflector and central region after installation finished

Design of Ion Source Magnet Distribution of CRM cyclotron for CYCIAE-100

JIA Xian-lu, ZHANG Tian-jue, ZHONG Jun-qing, CHEN Rong-fan, LU Yin-longYAO Hong-juan, SONG Guo-fang

In order to have a good command of H-multicusp source technology, a new source of 15-20 mA will been designed during the 11th five year plan. To reach the goal, the source will be designed on the basis of the previous 10 mA ion source at CIAE and reference on the source at Canadian National Laboratory TRIUMF. This paper gives a description on the distribution of the cusp source.

1 Source describtionH-ions in the multicusp are produced by thermal agitation of H atoms or H2 molecules in contact

with hot energetic electrons generated by the negative bias of the filament. The most effective way of increasing the impact is the proper use of magnetic field, and all kinds of plasmas will be restricted by the magnet field. To increase the density of the plasma around the central axis and to restrict effectively the plasma, we adopt the multicusp H-magnets structure, as is shown in Fig.1. The field is composed of radial magnet and azimuth magnet. The section of the radial magnet is 10 mm×18 mm, and it is 13 mm×18 mm for the azimuth magnet. The two kinds of magnets are placed alternately.

The multicusp H-magnetic field is calculated by the 2 dimensional software POISSON, and the result is shown in Fig. 2 indicating the magnetic field distribution. To verify the result, we also calculate the magnetic field using the 3 dimensional software ANSYTS. In the calculation the magnetic induction is Br=1.1 T, and the relative magnetic permeability is 1.07. During the magnetic field calculation, the radius of the air is 0.2 m, and its height is －0.101 6-0.304 8 m. The result is compared with that by the 2 dimensional calculation, as is shown in Fig. 3, in which the magnetic field distribution along the radial direction at the 0° and 18° respectively. From the figure, we can see that the results basically agree with each other, though it is lower for the 3 dimensional calculation than the result from 2 dimensional calculation. The main reason for this difference is that during the 3 dimensional calculation, the leakage field is taken into consideration. Therefore, the result by 3 dimensional calculation is more reasonable.


Fig. 1 Multicusp magnet arrangement

Fig. 2 Two dimensional magnetic field on 0°

Fig. 3 Result comparison of 2 dimension and 3 dimension

2 Filter magnetIn order to filter the fast electron, there are water cooled dipole columns near the plasma exit

aperture. The H-multicusp source of CYCIAE-30 used this kind of filter to separate the electron and H-ion. The principle of this kind of filter is that the magnetic rigidity of electron is much less than that of the H-ion, so the electron is deflected easily by magnetic field, and as a result the H- ion can go through the field.

Nowadays this method is not widely used, and instead we take use of virtual filter to filter the fast electron. A typical method for this is that the top azimuth magnet is taken out and the top bar of the centre


line “N” polarity is replaced by an “S” bar, so there is a 50-80 Gs magnetic field in the central region of the cavity. In addition, the extraction lenses will be installed to this dipole field so that the mixed field is about 150 Gs. We refer to this as version O, as is seen in Fig.4 (up). The 10 mA ion source of CIAE adopts this type of mode.

Due to the fact that a simple virtual filter is either too thick that the slow electrons are also filtered or too weak that it does not filter the fast electrons completely when the arc power is high, it is important that the filter should be thin and have proper strength. For this reason, a more elaborate arrangement of virtual magnet bars was made as is shown in Fig. 4 (down). Base on the 10 mA ion source, the top bar was replaced by a “S-N” combination and two equal size but shorter S bars in each side, and a pair of antidipole (small N bars ) was added in the indicated location, referred to as version E. The magnetic field arrangement on the centre line of version O and version E is shown in Fig. 5.

Fig. 4 Version O virtual filter (up) and version E virtual filter (down)

Fig. 5 Magnetic field arrangement on the centre line of version O and version E

Real line——version O; Broken line——version E

Research on Quadrupoles Along Injection Line for CRM Cyclotron of CYCIAE-100


WEI Su-min, LU Yin-long, ZHANG Tian-jue, LI Zhen-guo, WU Long-cheng, WANG Zhen-hui, Ge Tao, CAI Hong-ru, PAN Gao-feng, YANG Jian-jun

According the matching result of the beam transport along the injection line for the CRM Cyclotron of CYCIAE-100, a pair of quadrupole should be placed along the injection line in order to match the beam to the entrance of the deflector in the central region. Based on the design, the maximum magnetic field gradient for this pair of quadrupole lenses is 21.3 Gauss/mm, the effective length is 120 mm, and the radium of the inner aperture is 50 mm. Besides, it is required that the geometry structure of the quadrupole can be rotated by ±5° about the axial line.

Since this pair of quads are positioned in the inner part of the injection line of the main magnet for the 100 MeV compact cyclotron, the space is fairly limited and there is only the place with 120 mm reserved for their installation, the most difficult part for the design is how to make the best use of the limited space to arrange a reasonable layout of the magnet structure and coils in order to meet the requirement of magnetic field gradient.

To make the fabrication convenient and meanwhile further obtain the symmetry, the circular pole is chosen to substitute the hyperbola pole shape in theory. The inscribed circle radius of the sector is 25 mm, the inner radius of the return yoke is 54 mm and the outer radius is 59 mm, the mechanical length is 100 mm. A quarter of the profile of the quadrupole is shown in Fig. 1. According to the calculation, when the number of excitation ampere-turns is 660 A, the magnetic field gradient will be 26.7 Gauss/mm, which meets the design requirement. The power supply design will be based on the data. From 0 to 20 mm along the radius, the uniformity of the magnetic field gradient is 0.005. In the case that the mechanical length is 100 mm in 3D simulation, the effective length of the quadrupole is 121.8 mm. It meets the design requirement.

Fig. 1 Profile of quadrupole

Due to the limit of space, we meet a number of technical difficulties on the winding during the fabrication of this pair of quads. In initial design, the wires with inner water cooling of 6 mm×6 mm were be chose to wind the coils, and the outer surface of wire was insulated by glass tape. For each sector, there are 6 turns of wires. Since the space is limited and the hardness of the wires is high and the insulation layer is thin, the insulation can be easily destroyed during the winding. As a result, the coils can not be


winded. In order to solve the problem, we selected the lacquered wire of 1.8 mm to replace the inner water cooling wire of 6 mm×6 mm. For each sector, there are 15 turns of wires. In each coil, a 6 mm×6 mm inner water cooling coil is inserted as the water cooling tube. The experimental result shows that the tight contact between lacquered wire and the inner cooling tube is impossible, so the cooling effect is not ideal and the temperature increase of the coil is relatively obvious. Later the capillary water cooling wire with 2 mm in outer radius and 1 mm in inner radius is used to do the winding. The polyimide film is used outside for insulation. For each sector, there are 15 turns of wires. Since the wires are relatively close to each other and the insulation layer is thin, the insulation layer can be easily destroyed during the installation and thus result in short circuit. In the end, after extensive investigation and experimentation, the magnet adopts a divided structure, and it is winded by water cooling tube with 3 mm in outer radius and 2 mm in inner radius. The glass tape is used for insulation and for each sector, there are 11 turns of wires. Such a design can substantially reduce the difficulty on winding, and thus effectively solve the problem that the insulation layer of the wire is vulnerable. In the meantime, the water cooling effect is fairly reasonable.

Fig. 2 Picture of fabricated quads

Fig. 3 Result comparison of magnetic field

■——Measured value；●——Calculated value

Currently the fabrication of this pair of quads has been finished and the magnetic field mapping has been accomplished. The fabricated quads are shown in Fig. 2. The comparison between the theoretical value of the magnetic field and the real mapping value is shown in Fig. 3, from which it is easy to conclude that the mapping result is close to the theoretical value. The pair of quads will be used in the beam development along the injection line of the CRM Cyclotron for CYCIAE-100 and replace the normally used solenoidal in the past to take the role of axial focusing. It will be mainly used to meet the


injection optical requirement for the deflector matching.

Mapping System, Magnetic Measurement and Shimming in CRM of 100 MeV Cyclotron

ZHONG Jun-qing, LU Yin-long, YIN Zhi-guo, CHU Cheng-jie, ZHANG Tian-jue, WANG Chuan, WANG Zhen-hui, LIN Jun, LI Zhen-guo, GE Tao, WU Long-cheng,

YAO Hong-juan, JIA Xian-lu, WEI Su-min, XING Jian-sheng, YANG Jian-jun, ZOU Jian

The CRM is a compact Cyclotron accelerating H－, in view of the fact that there is a wide range of magnetic field in median plane of the machine, as the only method to check the magnetic field distribution, magnetic measurement is required to have a high accuracy, stability and repeatability. Based on the mapping system of CYCIAE-30, we designed a new one to meet the requirements of mapping and shimming of the CRM, as is shown in Fig. 1.

Fig. 1 Mapping system of CRM

1 Improvement on the mapping system in the CRMIn order to perform automatically the measurement and data acquisition of the magnetic field,

considering the field distribution and high accuracy requirement to reduce the error of the system, we selected the Group 3 Digital Tesla-meter and LPT-141 hall probe with 0.01% precision in our System. The hall probe was fixed in rotatable beam, the material of which was fibre-glass epoxy, It is required that the probe should be kept moving in the median plane of the machine.

Apart from the probe precision, the other major factor that determines the error of the mapping system is the precision of the mechanism system. The mapping system of the CRM has been improved


substantially compare to that of CYCIAE-30, including the separation of driven shaft with orientation shaft, for which the driven shaft was fixed above the orientation shaft. The orientation shaft was tightly fastened with beam by dowel and the angle encoder was fixed to the bottom of the orientation shaft. In this way we can get the rotating angle of the beam by encoder. Based on the data from the encoder, the computer controls the stepping motor to rotate the beam. From the result of mapping, it can be seen that the precision of the mapping system in angle orientation was improved 10 times to the old one.

2 Measurement and shimming of the magnetic field in the CRMThe measurement and shimming of the magnetic field in cyclotron was a process that requires

several times of repeating. During the magnetic field measurement for the CRM, we need to determine the center of the field by a fine adjusting of the hall probe position before the magnetic field measurement can be conducted. We must check and estimate the rationality of field distribution shown in screen by computer graphics method during the mapping, because there was a fairly large number of data concerning the magnetic field. After completing the mapping, we tracked the particle in field by DYNA2000 edited by Dr. Zhang. Consequently, we can obtain the maximum frequency, the mostly desired frequency, the minimum phase shift, the shift of isochronous field and the vertical betatron frequency, et al. Grounded on data of the isochronous field shift and the vertical betatron frequency, we got the data of the shimming bars by the relation between increment of the bars and that of the magnetic field. It has undergone 7 times of magnetic field measurement and 6 times of shimming before we enabled the main magnetic field match the requirement of the accelerated particle. The shimming of the first harmonic was included in the last bars machined. The shimming results of vertical betatron frequency and isochronous field are shown in Fig. 2 and Fig. 3 respectively.

Fig. 2 Shimming of vertical betatron frequency in CRM


Fig. 3 Shimming of isochronous field in CRM

Physic Design for Pulsed Beam System in CIAE’s CRM Cyclotron for CYCIAE-100

AN Shi-zhong, WU Long-cheng, ZHANG Tian-jue, YAO Hong-juan, ZHONG Jun-qing, JIA Xian-lu, GUAN Feng-ping, JI Bin, WEI Su-min

A test beam line for pulsed beam generation system for 10 MeV central region of model (CRM) compact cyclotron is being constructed at China Institute of Atomic Energy (CIAE). A 70 MHz continuous H-beam with the energy of several decades keV or a hundred keV will be pulsed to the pulse length of less than 10 ns with the repetition rate of 1-8 MHz.

The buncher and chopper systems will be the main parts in the beam pulsing system. A buncher with the frequency of 70.487 MHz will be used to compress the DC beam into the RF phase acceptance of ±30° of the CRM cyclotron. The corresponding pulse length is 2.36 ns for the RF phase acceptance of ±30°, which will meet the requirement of pulsed system. The design of the beam chopper will be the key part in the pulsing system. The sine waveform with the frequency of 2.2 MHz will be used in the chopper after many discussions.

Considering the limits of the space, the buncher and chopper systems will be put together in the pulsing system. This choice will be compact and the beam line is not long. The layout of the new beam line with the pulsing system is shown in Fig. 1. The symbols in the Fig. 1: FC-Faraday cup, SLIT-limiting slit for the chopper, EINZEL-Three-aperture Einzel lens, SOL-Solenoid, Q-Quadrupoles, STR-Steering magnets. This design adopts ESQQ focusing structure. The continuous H-beam with the energy of 40 keV from the multi-cusp ion source is focused into two long waists by the Einzel lens at first and the beam chopper and buncher are put at this waists. The chopper is put behind the buncher and both of them will be put the same vacuum chamber. The beam after the buncher and chopper will be transported into the inflector of CRM cyclotron by the solenoid and quadrupoles. There is a limiting slit at the entrance of inflector of CRM, it will be the common working slit for the chopper and buncher. Another auxiliary slit behind the solenoid will be used to test the chopper’s working state.


In the pulsing system, the main limits in the transverse direction are the beam envelope and the beam size at the entrance of inflector and in the longitudinal direction are the energy spread and beam current. The energy spread will affect the pulse length and the inject matching. The beam current not only affects the whole beam envelope and the beam size at the entrance of inflector, but also the bunching efficiency of the buncher. Both the buncher and chopper need the beam go through them with small waists and the beam size needs to be limited about 2.6×4 mm at the entrance of the inflector. In order to keep the bunching efficiency, the beam current is limited below 3.5 mA during the calculation results. So, considering the requirements to the pulsing system and inject matching, a beam limit device at the exit of the ion source is to used to limit the beam current and beam emittance.

The calculation for the beam optics includes 2 parts: one part is to fix the initial beam current to calculate the beam envelope and beam size with different initial emittance, which will find the optimal initial matching condition; another part is to fix the initial emittance to calculate the beam envelope and beam size with different beam current, which will get the beam current limiting in the transverse. The beam envelope is controlled under 15 mm in the design and the beam size at the entrance of the inflector attains the requirements for the injecting match. All of the results for the different beam current and initial emittance are shown in Figs. 1 and 2.

Fig. 1 Results for different initial emittance


Fig. 2 Results for the different beam current (Initial emittance 48π mm·mrad)

Design and Calculation of Buncher in Pulsing Beam System


1 Basic theory of the buncher The buncher system is the one of the main parts in the pulsing beam generation system for 10 MeV

central region of model (CRM) compact cyclotron and it will modulate the beam in the longitudinal direction. The beam with special phase width will be compressed into very short pulse in the time with buncher. With the modulation of a buncher, the charged particles in the continuous beam segments will come into being very short pulse after a long drift, and this pulse length should be less than the RF phase acceptance of the CRM cyclotron, that is, the pulsed beam can be captured by the RF system of the cyclotron completely. Because the velocities of the charged particles will be modulated with the buncher, an energy spread in the pulsed beam will be introduced. This energy spread affects not only the pulse length, but also the injection matching to the cyclotron. The drift in the buncher will affect the bunching voltage, which will determine the energy spread, and the space charge effects for different drifts will affect the bunching efficiency. So, for the design of buncher, the energy spread produced by buncher system and bunching efficiency affected by space charge effects will be studied in detail.

For the CRM cyclotron, the 70.487 MHz RF phase acceptance is ±30° and the corresponding pulse length is 2.36 ns. This pulse length will meet the requirement of pulsed system. The buncher will adopt the same frequency of 70.487 MHz. The mesh structure with single drift and dual gaps will be used in the bunhcer, see Fig. 1. The signs from RF amplifier will be added into inner meshes and outer meshes will be grounded. The silk in the mesh is made of gold-filled tungsten and the thickness of the silk is about 50-100 µm. The distance between the mesh is 5 mm, the distance between the silk is 3 mm, and the length of middle electrode is 20 mm (adjustable). The drift for the buncher is 1.384 m in the pulsing beam line with the length of 2.071 m in total. The bunching waveform is used sine form under non-relativity and the amplitude of the bunching voltage can be expressed:

m1(kV) 35.4

(MHz) (m)V

f L

The amplitude of the bunching voltage is Vm=0.363 kV. So Vm<1.5 kV is enough for the amplitude of the bunching voltage.

2 Simulation of the buncher Figure 2 shows the simulation results for the buncher. From the results, the bunching efficiency is

about 63.5% and the produced energy spread is 1.5% with the amplitude of 585 V. Figure 3 and Table 1 show the calculation results with different beam current. The bunching efficiency is reducing as the increasing beam current at the drift of 1.384 m. It reduces to 52% under the 3.5 mA beam current, and the energy spread is reduced to 1.1% at the same time. So, the acceptable beam current should be controlled under 3-3.5 mA.


Fig. 1 Structure of buncher

Fig. 2 Bunching results with buncher

Fig. 3 Simulation results with different beam current

Real line——Sine wave；◇——I=0 mA，=0.635；△——I=1 mA，=0.63；○——I=2 mA，=0.59；■——I=3 mA，=0.55；＋——I=3.5 mA，=0.52


Table 1 Amplitude Vm，maximal energy spread ΔE/E and bunching efficiency η

I/mA Vm/V η/% (ΔE/E)/%

0 585 63.5 1.5

1 670 63 1.4

2 698 59 1.5

3 750 55 1.4

3.5 780 52 1.1

Design of Buncher in Pulsing Beam System


For the buncher system with the sine waveform, the beam with phase width of 110 will be bunched in effect. So the pulse length got from beam chopper should be less than this phase width. In order to increase the effective beam current, the good bunching efficiency should be kept. The period for the buncher with frequency of 70.487 MHz is 14.18 ns and the pulse length should be less than 8.67 ns (corresponding the bunching phase width of 110) after chopper.

In the beam pulsing system for the CRM cyclotron, the design of the chopper will be the key part due to the very short pulse length (<10 ns), especially the choice of the chopping waveform. In order to get the pulsed beam with repetition of 1-8 MHz and pulse length of 8 ns, the sine waveform is used in the chopper. The frequency of the chopper is chosen 2.2 MHz, which is the 32-divided RF frequency of the buncher. We can get the pulsed beam with repetition of 4.405 MHz after the chopper.

The structure of beam chopper is shown in Fig. 1. The symbols in Fig. 1: l is the length of the plate, g is the gap between the plates, S is the drift distance between the chopper and working slit, r is the radius of the slit.

Fig. 1 Structure of beam chopper

For the chopper with sine waveform, the amplitude of the voltage can be expressed:

where E0 is the energy of particle, b=r+a, a is the half width of the beam at the slit, τc is the pulsed length.

From the above equation, in order to decrease the amplitude of the voltage, the length of plate or drift should be enlarged and the g or b should be decreased. So, it is better to put the chopper and the slit at the waist of the beam respectively.


Figure 2 shows the exact structure of the chopper and the parameters of the chopper are: the length of plate is 100 mm; the width of the plate is 50 mm; the thickness of the plate is 5 mm; the gap of g is 20 mm; the drift is 1.274 m; working RF frequency is 2.2 MHz with sine waveform. The pulsed length after chopping is τc=8 ns or ±3° in RF phase of chopper. The chopping waveform is shown in Fig. 3 and the amplitude of 4 kV is needed to keep the pulse length of 8 ns.

The average beam current after chopping is:

Where I0 is the DC beam current. If I0=1 mA, the average pulsed beam current is about 19.2 µA with the pulse length of 8.67 ns. That is, about 20 µA pulsed beam can be got for the 1 mA DC beam.

Fig. 2 Structure of the chopper

Fig. 3 Sine waveform for chopper


Design of Vacuum Test Stand

PAN Gao-feng, LI Zhen-guo, ZHANG Tian-jue, CHU Cheng-jie, WU Long-cheng

1 Purpose of vacuum test standThe vacuum test stand will be used to measure the outgas-rate of material surface, in order to search

some materials with less outgas and the techniques to improve the vacuum quality of the material surface.

2 Necessity of building a vacuum test standThe inner surface of the 100 MeV Cyclotron vacuum chamber is about 200 m2, and the possible

pumping speed is about 100 000 L/s. In order to keep the pressure in chamber less than 5×10 －6 Pa, the outgas-rate of vacuum chamber materials must be less than 2.5×10－7 Pa·L/(s·cm2).

3 Main consideritions of the design3.1 Method to measure the outgas-rate

The outgas-rate of the materials will be obtained by comparing the pressure in the Test stand’s chamber in two conditions, with and without the materials, when the pumping time is the same. The outgas of the materials will increase the pressure in the chamber.3.2 Pumping speed of the pumps used to the test stand

Based on the foundation P=Q/S (P is the pressure in the chamber of the Test Stand, Q is the gas flow into the chamber, and S is the pumping speed of the pumps used to the Test Stand). We can find out that if the pumping speed is smaller, the pressure is higher. So it is better to make the pumping speed smaller. A turbomolecular pump with pumping speed of 80 L/s is chosen as the main pump of the Test Stand.3.3 Size of the sample to be tested

Based on the foundation P=Q/S, we can also find out that the faster the gas flow is, the higher the pressure will be. To increase the test precision, we can increase the surface of the tested sample. In this case, the surface of the tested materials should be no less than 1 000 cm2.3.4 Ultra-pressure of the test stand

In this case, the pressure produced by the outgas of the tested materials is 3.125×10－6 Pa. To increase the test precision, the pressure should be lower when the test materials are absent. As a result, it will be better if the ultra-pressure of the test Stand can be lower than 2×10－6 Pa.

To keep the pressure in the chamber lower than 2×10－6 Pa, the materials of the seals used in the Test Stand should be metal and the chamber should be baked.

To know the component of outgas from the tested materials, a Residual Gas Analyzer System is necessary.

Progress Report of ISOL System of BRIF Project in 2006

Nuclear Physics Division ISOL Group

In 2006 the ISOL system was optimized based on previous work, a revised feasibility report and a


primary design report was finished. Also the investigations on some key technical issues and detail design of some key parts of the ISOL system were carried out. The details of these works are described as following.

Considerable investigations on maintains of target ion source have been carried out in 2006. The issues such as installation, relation with the construction and activation of the shielding iron were investigated. A combination shielding plate of 15 cm thick iron and 5 cm lead plate is adapted to shield the radiation from target so that the operator can disassembly the connections safely. The shielding plate is designed so that all the movable parts can be maintained even the target ion source is high level radioactive.

Some investigations related to dipole magnet was also carried out. A α and β correction coils was manufactured and tested on a dipole magnet. With α=5.25×10 － 3 and β=0.276 the error between calculation value and measured value is 2% and 6% respectively, the results show good consistency. Also the effect of permeability of vacuum chamber material in dipole magnet on magnetic uniformity is investigated. The possibility which uses Ti as vacuum chamber material is explored. The results are useful for the selection of vacuum chamber. Another test which helps to determinate the ripple requirements of dipole power supply is carried out. In this test the magnetic fluctuations of dipole magnetic field caused by different level ripple of power supply is measured.

The collection system of radioactive gaseous waste from vacuum pumps nearby the target ion source was designed, which will collect and store the radioactive gases until the radiation level decay to allowed level, in normal operation condition this collection system can collect the gas from pump of target area for more than half year. A test which put seal material used by turbo pump into radiation field is carried out, the results show that it is promising for the turbo pump to operate for a reasonable period.

The design of control system of ISOL is in progress. The PLC architecture is adopted to control the system. Several manufactures have been contacted as potential vender.

Progress Report of Superconducting Booster of Beijing Radioactivity Ion-Beam Facility in 2006

Superconducting Booster Group of BRIF

The superconducting booster is a main part of the Beijing Radioactivity Ion-beam Facility project (BRIF). As the designed target of the superconducting booster, an energy gain 2 MeV/q for β=0.118 ions will be expected. Up to date, we have completed all the design work. The development for some components is going smoothly.

Four Niobium-Copper quarter wave resonators cavities will be adopted in the superconducting booster. Working frequency is 150.4 MHz and geometry factor is 26.17 . The technology process to manufacture the copper base cavities has been determined, and the machining contract with the manufactory has been signed. The manufacturer has made a bulk copper into a mold and will finally process. We launched on the sputtering technology in 2006, and got a better sputtering process after researched, studied and tested.

The cryostat is mainly composed of a vacuum container, a heat insulation shielding shell filled with liquid nitrogen, a liquid helium vessel and support structure to install cavities. The static power consumption is less than 10 W. The vacuum is in 10－5 Pa range at room temperature. The rate of leak gas


is less than 1×10 － 9 m3·Pa/s. The manufacturer has made all components and begins to assemble the cryostat. We will check it in 2007.

The radio frequency system provides RF power, base phase and base time for the superconducting booster and low energy bunching system. The working frequency of the booster is 150.4 MHz and the fundamental frequency of the bunching system is 6 MHz. The phase stabilization system will connect them together. Now we are ready for invitation of bids. In the field of control system, we bought a set of development platform which uses VME-bus construct as the hardware and uses VxWorks as the operation system. It will be used to develop the control system based on EPICS (Experimental Physics and Industrial Control System, EPICS).

We had got a good progress in design and development of the key components of the superconducting booster in 2006. Next year we will build clean room, the cryostat, the sputtering platform, the RF-instruments and the bunching system. In a word, the superconducting booster is going smoothly as our schedule.

Upgrade of Tandem Accelerator Injector System

BAO Yi-wen, FAN Hong-sheng, GUAN Xia-ling, HU Yue-ming, HUANG Qing-hua, KAN Zhao-xin, SU Sheng-yong, WANG Xiao-fei, YANG Bing-fan, YANG Tao, YOU Qu-bo

The construction and the commissioning of upgrade program of Beijing HI-13 Tandem Accelerator Injector System is basically completed at Beijing Tandem Accelerator National Laboratory (BTANL) in 2006. It has put into normally operation since October of 2006 and will be formal accepted by BRIF on October of 2007.

This project is commonality required and supported by tandem nuclear national laboratory, Beijing radioactive Ion-beam Facility project and the Object of AMS of Nuclear division.

The new injector system include two beam lines, One is high voltage platform beam line for normal operation, The platform voltage promoted form 150 kV to 300 kV to match the superconducting boostor. The other is high resolution achromatic beam line that consists of an electrostatic and a magnetic analyzer specially for Accelerator Mass Spectrometer (AMS). The design parameter of mass resolution should be better than 380.

Fig. 1 Mass spectrometer of HfO2－ at image slit of magnet analyzer


The good results of both stably working voltage of the platform and the mass resolution are achieved. Figure 1 is the typical scan spectrum over isotopes of HfO2

－ beam from new injector. It is shown the mass resolution better than 480.

Figure 2 is a picture of the new beam ling of the upgrading injector system of HI-13 tandem accelerator. It can be find the new accelerating tube, the quadrupole lenses, the electrostatic analyzer and the dipole analyzer.

Fig. 2 Picture of new beam ling of upgrading injector system

Primary Shielding Design for BRIF

WU Hai-cheng, TANG Hong-qing, SHI Yuan-ping1

(1 The Fourth Institute of Nuclear Engineering)

1 Introduction1.1 Design principle

The radioprotection of the Beijing Radioactive Ion-beam Facility (BRIF) has been conducted under the instruction of GB 18871—2002[1]. The shielding of the facility has designed based on the value listed in Table 1.

For the sake of management, the whole building is classified into four areas: 1) General Area, registered radiation workers can enter freely; 2) Supervised Area, personal dose monitoring is expected and no other special protection method or

procedure are required; annual equivalent dose of a worker in this area will not be greater than 5 mSv;3) Controlled Area I, access to this area is limited and with at least 72 h passed after cyclotron

stopped; permissions and the entering procedures are required; 4) Controlled Area II, targets are disassembled here and access to this area is strictly limited. Only


when transferring radiation wastes and with the permission and company of a radiation protection officer, the registered radiation workers can enter and stay after the target has been put into lead pot and the room been decontaminated.

Table 1 Design value for radiation shieldingArea Design value National standard

Public Area <0.01 mSv/a

Site Boundary <1 mSv/a

General Area <0.5 μSv/h

Supervised Area <2.5 μSv/h 1-5 mSv/a

Controlled Area I <10 μSv/h >5 mSv/a

Controlled Area II <100 μSv/h <20 mSv/a

Ground Water Activation <11 mSv/h

1.2 Evaluation of radiationThe BRIF contents several radiation sources: beam loss at the Beam Extraction Area (BEA) of the

cyclotron, Scattering Loss during beam Transport (SLT), targets of Isotope Separation On-Line (ISOL), Isotope Production (IP), White Light Neutron Source (WLNS), Quasi Mono-Energetic Neutron Source (QMENS), Strong Neutron Source (SNS), Radiation Biomedicine (RB) and Radiation Effect (RE). In the design, Heavy Concrete (HC, density=3.3 g/cm3), ferroconcrete (OC, density=2.42 g/cm3), iron (Fe), copper (Cu) and lead (Pb) are used as materials for bulk or partial shielding.

During the design, the dose distribution of different source in different target station or lobby is calculated separately and then the dose rates induced by different radiation source are spliced together to evaluate the radiation dose caused by the whole facility.

At the first stage, yields, energy spectra, angular flux of the projected neutron and gamma when protons bombard on an assumed cylinder target are calculated and serve as source term of the next step. At the second stage, a spherical shielding model is used to estimate the thickness required of bulk shielding for a certain radiation source. At the third stage, a 3-D geometric structure of a target station or lobby is obtained, and Monte-Carlo simulation is performed to calculate dose distribution. The angle of proton incidence, the scattering effects of sky and ground are considered in the 3-D calculation. At the fourth stage, the dose rates induced by the whole facility are evaluated based on dose rates on the edge of different target stations and lobbies. The contribution from different radiation sources are spliced together with the location of the source and the working mode of the cyclotron taken into account, and obtain dose rates on the boundary of the facility. At the final stage, small adjustments are made to keep dose rates on the site boundary under the limits.

When converting flux to dose rate, ICRP-74[2] neutron flux-to-dose rate conversion factor and ICRP-21[3] photon flux-to-dose rate are used. According national standard, a factor of 2 has been employed as safe coefficient and another factor of 2 been used to correct calculated results.

2 Shielding design for BRIF2.1 Beam loss assumption

The radiation sources assumed are listed in Table 2. All the target stations are assumed to be working at the strongest beam loss condition that probably can be obtained during the annual operation time.


Table 2 Assumed beam loss of BRIFLocation Name Target material Target thickness/cm Energy/MeV Intensity/μA Annual operation time/h

Cyclotron Lobby Northern BEA Al 4 100 20 1 980

Southern BEA Al 4 100 20 1 320

Loss on line Al 4 100 10 3 300

ISOL target station ISOL target Pb 1.6 100 90 1 650

IP target station IP target Pb 1.6 75 50 330

Large Neutron

Experiment Lobby

QMENS Li 0.03 100 10 330

Beam dump

of QMENSC 4.1 100 10 330

WLNS Pb 1.6 100 20 330

SNS Pb 1.6 100 200 330

RE Fe 4 100 1.0×10－3 264

RB Pb 1.6 100 3.0×10－1 66

2.2 Radiation shieldingA semi-underground structure is proposed in construction the BRIF. The floor board of the first floor

will be on －4.0 m and constructed by 1.0 m ferroconcrete and 0.5 m ordinary concrete. Table 3 gives the shielding requirements for the target stations and cyclotron lobby. On the direction of high, different shielding material and thickness may be used when the shielding effect of soil is considered.

Table 3 Required shielding material and thickness for BRIF

Location Name Partial shielding/m Eastern wall/mSouthern

wall/mWestern wall/m Northern wall/m Top /m

Cyclotron

Lobby

S/N

BEA0.3 Pb

3.0HC[－4,8.5]1)

3.6OC[－6.8, －4]

3.0HC[－4,8.5]

3.6OC[－6.8, －4]

3.6OC[－4,8.5]3.0HC[－1,8.5]

3.0OC[－4,-1]2.5OC

SLT

3.0HC[－4,8.5]

3.6OC[－6.8, －4]

3.0HC[－4,8.5]

3.6OC[－6.8, －4]

3.6OC[－4,8.5]3.0HC[－1,8.5]

3.0OC[－4, －1]2.5OC

ISOL target

stationISOL 3.5HC[－4,1]

3.0HC[－1,1]

3.0OC[－4,-1]2.5HC[－4,1] 3.0 HC[－4,1] 2.6HC

IP target

stationIP 0.3Cu

2.5HC[－4, －1]

3.0OC[－4, －1] 2.0HC[－4, －1] 2.6HC[－4, －1] 2.0HC

Large

Neutron

Experiment

Lobby

QMENS 1.9OC[－4,5] 1.9OC[－4,5] 1.9OC[－4,5]3.0HC[－4,8.5]

3.6OC[－6.8, －4]1.4OC

Beam dump

of QMENS

1.4Fe

+0.4HC1.9OC[－4,5] 1.9OC[－4,5] 1.9OC[－4,5]

3.0HC[－4,8.5]

3.6OC[－6.8, －4]1.4OC

WLNS/SNS0.3Pb

+1.8HC1.9OC[－4,5] 1.9OC[－4,5] 1.9OC[－4,5]

3.0HC[－4,8.5]

3.6OC[－6.8, －4]1.4OC

RE/RB 1.5HC 1.9OC[－4,5] 1.9OC[－4,5] 1.9OC[－4,5] 3.0HC[－4,8.5] 1.4OC


3.6OC[－6.8, －4]

Note:1) use 3.0m HC from -4 m to 8.5 m high

Reference:[1] GB 18871—2002. The basic standards for protection against ionizing radiation and for the safety of radiation sources.

National Standard, 2002.

[2] ICRP Publication 74. Conversion coefficients for use in radiological protection against external radiation. Annals of the

ICRP, 1996, 26: 3-4.

[3] ICRP Publication 21. Data for protection against ionizing radiation from external sources. Annals of the ICRP, 1973.

Documents

IMPORTANT NUCLEAR SCIENCE ENGINEERING · Web viewReport of Beijing Radioactive Ion-B. eam. Facilities (BRIF) in 2006. Engineering Division of BRIF Project. Beijing Radioactive Ion-beam