MAX (Contract Number )

Transcription

1 MAX (Contract Number ) DELIVERABLE Number MHz vs. 176 MHz Injector Comparison & Choice Authors: Chuan Zhang, Horst Klein, Dominik Mäder, Holger Podlech, Ulrich Ratzinger, Alwin Schempp, Rudolf Tiede, Markus Vossberg First reporting period: 01/02/ /07/2012 Date of issue of this report: 31/01/2012 Start date of project : 01/02/2011 Duration : 36 Months

2 DISTRIBUTION LIST Name BAYLAC Maud, CNRS BIARROTTE Jean Luc, CNRS BOULY Frédéric, CNRS/TED BOUSSON Sébastien, CNRS BRUCKER Romain, EA DARGES Bernard, TED DE GERSEM Herbert, KUL ESSABAA Saïd, CNRS FERNANDEZ RAMOS Pedro, EA GARBIL Roger, EC GARDES Daniel, CNRS JUNQUERA Tomas, ACS KALININE Amélie, CNRS KLEIN Horst, IAP MARTIN SANCHEZ Juan, ADEX MASSCHAELE Bert, KUL NEVADO Antonio, ADEX PERROT Luc, CNRS PIERINI Paolo, INFN PIRES Rui, FE UCP PODLECH Holger, IAP ROGGEN Toon, KUL SAUGNAC Hervé, CNRS SIERRA Serge, TED SPYROU Smaragda, CNRS URIOT Didier, CEA VANDEPLASSCHE Dirk, SCK CEN ZHANG Chuan, IAP Comments E copy of the document [MAX] (D 2.1) 352 MHz vs. 176 MHz Injector Comparison & Choice Dissemination level: PU Date of issue of this report: 31/01/2012

3 TABLE OF CONTENTS 1. INTRODUCTION FROM EUROTRANS TO MAX: NEW DESIGN CONCEPTS FOR THE INJECTOR MHz vs. 176 MHz Other Major Changes THE RFQ ACCELERATOR Design and Simulation Results Simulation Based on the LEBT Output Distributions Error Studies THE CH DTL PART Design and Simulation Results Error Studies CONCLUSIONS REFERENCES

4 1. INTRODUCTION Launched by the European Commission in 2005 and ended in 2010, EUROTRANS [1] was a EUROpean research programme for the TRANSmutation of high level nuclear waste in an accelerator driven system. As a successor of the EUROTRANS project, MAX [2], the so called MYRRHA Accelerator experiment research and development programme, has been started in 2011 and will continue until Different than EUROTRANS which was a pure research project, MAX is pursuing not only to continue the R&D studies but also to deliver an updated consolidated design for the real construction including prototyping and demonstration in Mol, Belgium. Table 1: Specifications of the required proton beams for EUROTRANS & MAX. Parameter EUROTRANS XT ADS EUROTRANS EFIT MAX Operation (Design) intensity [ma] ( 5 ) 20 ( 30 ) ( 5 ) Output energy [MeV] Beam trip number >1s: <5 per 3 month operation cycle >1s: <3 per year >3s: <10 per 3 month operation cycle Beam stability (on target) Energy: ± 1 %, Intensity: ± 2 %, Beam Size: ± 10 % Beam time structure CW, with 200μs zero current holes at 1 Hz repetition frequency The specifications of the required proton beams for EUROTRANS (including both XT ADS and EFIT phases) & MAX are listed in Table 1, where the most demanding requirement is that the beam trips (i.e. the beam interruptions on the target) with long duration periods (in the order of second) have been restricted to very small amounts, because such beam trips will cause serious thermal stress and fatal damages to the sub critical core. These beam trip limits are two or three orders of magnitude lower than typical values found with existing 4

5 [MAX] (D 2.1) 352 MHz vs. 176 MHz Injector Comparison & Choice accelerators [3], so the primary concern for the design of the EUROTRANS or MAX driver linac is how to ensure such extremely high reliability. In the table it s also seen that, except the slightly different beam trip limit all other beam specifications for MAX are identical to those for XT ADS. Therefore, it has been decided that the layout of the driver linac for MAX will follow the reference design made for the XT ADS phase of the EUROTRANS project. Fig. 1 shows the schematic plots of the driver linacs for both EUROTRANS and MAX. It can be seen that the required MAX accelerator is very similar to that for the XT ADS phase of EUROTRANS, except the linac front end. Figure 1: The driver linac layout for EUROTRANS and MAX. During the EUROTRANS project, a 352MHz, 17MeV, and upgradeable 5 30mA injector, which consists of one RFQ accelerator, two RT (room temperature) CH (Cross bar H mode) DTL (Drift Tube Linac) cavities, and four SC (superconducting) CH DTL cavities, was designed and successfully accepted as the reference design by the project [4]. As shown in Fig. 1, the MAX injector will use the basic layout design from the EUROTRANS one, but some new design strategies and approaches, e.g. different resonant frequency and different type of the RFQ structure, have been proposed and applied to reach a more reliable CW operation at reduced costs [5]. In this deliverable, the 352MHz injector designed for EUROTRANS and the 176MHz one for MAX are compared in detail. 5

6 2. FROM EUROTRANS TO MAX: NEW DESIGN CONCEPTS FOR THE INJECTOR From EUROTRANS to MAX, some requirements have been changed. Cooled by liquid Pb Bi eutectic, the MYRRHA reactor will have a thermal power of ~80MWth in the ADS mode. The core geometry is optimized for an impinging proton beam energy of 600 MeV. Based on these data, it is found that the required beam intensity varies between 2.5 and 4 ma, depending on the burnup of the nuclear fuel [6]. Therefore, only 5mA will be taken as the design beam intensity, and the higher design intensity, 30mA, is not an option any more. Besides, the main differences for the linac front end are [5]: 1) The resonant frequency was lowered by a factor of 2, i.e. from 352MHz to 176MHz. 2) The 4 vane RFQ structure is now replaced by the 4 rod one. 3) The input and output energies of the RFQ were reduced from 0.05MeV and 3MeV to 0.03MeV and 1.5MeV, respectively. 4) The transition energy between the warm CH DTL part and the cold one has been accordingly dropped to 3.5MeV MHz vs. 176 MHz The most important change in the injector design is that the resonant frequency is lowered from 352MHz to 176MHz. The considerations for this change are as follows: The main point is the fact, as indicated in Fig. 2, that the shunt impedance of an RFQ, R p, is roughly proportional to f 1.5. Therefore, a major advantage for adopting a half resonant frequency is that the RF power consumption can be considably reduced. The lower frequency also enables the use of the 4 rod RFQ structure instead of the originally proposed 4 vane one. Fig. 3 compares these two kinds of mainstream RFQ resonant structures. The 4 vane structure works in the TE mode and its RF properties are determined not only by the vanes but also by the cavity wall, while the 4 rod one is actually a chain of λ/4 resonators and its inner structure is almost independent to the cavity wall. The pros and cons of the 4 vane RFQ are that it has relatively even RF power density and could be easily cooled, but it will have a large radial size at frequencies 200MHz and the construction and tuning are relatively complicated and expensive due to very tight tolerances. In case of the 4 rod RFQ, it could always have a compact radial size and an easy construction, tuning, 6

7 and even repair, but its local RF power density is typically ~2 times higher. At frequencies higher than 300MHz, the 4 vane structure is certainly the best choice for CW operation, but at 176MHz, the 4 rod structure is more attractive. Figure 2: A survey of R p values for RFQ accelerators (the original plot with the data marked in black is from [7]; in [8], the data marked in blue was added; the data marked in green are newly added, where the value for the IFMIF EVEDA RFQ was kindly provided by Dr. A. Pisent). Figure 3: RFQ resonant structures. 7

8 2.2. Other Major Changes As the test facility in Mol will be operated with beam intensities up to 4 ma, only 5mA will be taken as the design intensity. Consequently, the inter vane voltage could have a drop from 65kV to 40kV in order to further reduce the RF power per length by ~40%. The length of the RFQ could be kept constant by lowering the input and output energies by 40% and 50%, respectively. The 4m long structure which allows to use only one tank is in principle similar to the SARAF RFQ [9] which can be operated in CW mode with much higher power (more than 180kW) than we need for the MAX RFQ. For the CH DTL, the input energy is now lowered from 3MeV to 1.5MeV. Though it brings some difficulties to the beam dynamics design, it is favorable from the cavity design point of view: 1) The effective shunt impedance Z eff which is roughly proportional to ß 1 (see Fig. 4) is increased by ~30%, which saves RF power as well as makes the cooling easier. 2) It could compensate the cell length growth caused by the lowered frequency. Actually, the new frequency is also helpful for the CH cavity design. For example, the first cell of the first RT CH was lengthened from 3.4cm to 4.8cm, which provides more space for the field flatness tuning. Figure 4: RF power efficiency of multi cell structures [10, 11]. 8

9 As in the EUROTRANS case, the two RT CHs will also cover an energy gain of 2MeV, but both at 34% lower accelerating gradients E a for further reducing RF power. In the new design, the triplets have been moved further into the cavities (85mm instead of 35mm), which will not only lead to a better field flatness but also save the drift space. In total, this part will be still maintained compact. The four SC CHs have been decided to keep working at E a 4MV/m. They will take over some additional energy gain which was cut in the RFQ, so the total length will be somewhat longer. However, only a 5mA beam will be fed into the MAX injector, so some focusing elements from the previous design e.g. the 2 nd rebuncher cavity and two solenoids could be removed. Totally, the whole CH DTL part is even shorter. Moreover, the new SC CHs have much less gaps, which makes the construction work (e.g. welding) easier and cheaper. 3. THE RFQ ACCELERATOR Same to the EUROTRANS case, the beam dynamics design of the MAX RFQ was based on the New Four Section Procedure (NFSP) [12, 13], an efficient design method for modern RFQs, while the beam transport simulation was also performed with the PARMTEQM code [14] using 10 5 input macro particles Design and Simulation Results In Table 2, the detailed design and simulation results at 5mA for both the EUROTRANS RFQ and the MAX RFQ are presented. For a good comparison, the corresponding parameters of the SARAF RFQ for accelerating protons are also listed in the table. Obviously, from EUROTRANS to MAX, the transmission and transverse output emittances are still very similar, but the longitudinal output emittance is decreased considerably. In addition, the Kilpatrick factor is now only 1, well below 1.8, a safe value proven by the LEDA RFQ for CW operation [15]. And the minimum gap between electrodes is enlarged by 1mm. Both results are favorable for leading to a very reliable CW operation. Generally speaking, the MAX RFQ has quite similar main RF and geometric parameters to the built SARAF RFQ, so the R p value measured from the latter, 67 kωm, could be used as a reference and we could easily estimate 9

10 that the RF power of the former is 23.5kW/m. The SARAF experiments have shown that for one 8 hour CW operation at 40kW/m only two beam trips (in the order of ms) happened and the reached power record for CW operation is 50kW/m [16]. Therefore, it s clear that the MAX RFQ design is very reliable. Comparing the EUROTRANS 4 vane RFQ with the MAX 4 rod RFQ, it should be aware of course that the lower frequency would also be a corresponding advantage for the 4 vane RFQ, but would require a very bulky, heavy, and expensive cavity. Table 2: RFQ parameters for EUROTRANS & MAX. Parameter EUROTRANS@5mA MAX SARAF (H + ) f [MHz] I [ma] W in / W out [MeV] 0.05 / / / 1.5 U [kv] E s, max / E k a min [mm] m max g min [mm] t., n., ε rms in [π mm mrad] t., n., ε rms out [π mm mrad] 0.21 / / * / 0.19* l., ε rms out [π kev deg] * L [m] T [%] ~100 ~ * T 10mA [%] ~100 ~ * R p [kωm] 61 (MWS) 67 (after SARAF) 67 (meas.) P c [kw] 300 (MWS, +20%) * Simulated by A. Bechtold using the RFQSim code (no image effects or multipole effects). 10

11 At 5mA, the beam tranport plots for both the EUROTRANS RFQ and the MAX RFQ are compared in Fig. 5. Figure 5: Beam transport plots of the EUROTRANS (top) and MAX (bottom) RFQs. 11

12 3.2. Simulation Based on the LEBT Output Distributions Fig. 6 shows two versions of the LEBT (Low Energy Beam Transport) section designed by J. L. Biarrotte for the MAX project [17]. Figure 6: Schematic layouts of the MAX LEBT (top: short version, bottom: long version) [17]. For the design of the MAX RFQ, a 4D Waterbag distribution was used. By taking the output particle distributions from both short and long LEBT versions as the input distributions, new RFQ simulations were performed. Fig. 7 and Fig. 8 show that the input and output distributions at the beginning and at the exit of the RFQ for the different cases, respectively. 12

13 Figure 7: RFQ simulation results based on the short LEBT (top: transient, bottom: nominal). 13

14 Figure 8: RFQ simulation results based on the long LEBT (top: transient, bottom: nominal). 14

15 Table 3 summarizes the beam performance with different input distributions. It can be seen that the beam loss situation and the transverse emittance growths are still very satisfying in all cases. Induced by the wing form halo particles from both LEBT designs, all output longitudinal emittances are somewhat bigger, but still acceptable. Table 3: RFQ beam performance with different input distributions. Parameter 4D Waterbag Short LEBT transient Short LEBT nominal Long LEBT transient Long LEBT nominal t., n., ε rms in [π mm mrad] x., n., ε rms out [π mm mrad] y., n., ε rms out [π mm mrad] l., ε rms out [π kev deg] T [%] Error Studies The error studies have been carried out for the MAX RFQ with respect to seven input parameters: the intensity, emittance, inter vane voltage, Twiss parameters, energy spread, and spatial displacement, respectively. Table 4 gives the error settings, while Fig. 9 shows the lowest transmission is higher than 97% in all tested cases. Table 4: Error study settings and results of the MAX RFQ. Parameter Start value End value Design value Step length T min [%] I in [ma] t., ε un. in [π cm rad] U [%] Twiss α Twiss ß [cm/rad] ΔW [%] ~100 δx [mm]

16 Figure 9: Transmission as a function of test step for different input parameters. 4. THE CH DTL PART Same to the EUROTRANS case, the beam dynamics design of the MAX CH DTL was based on the KONUS method [18], while the beam transport simulation was also performed with the LORASR code [19]. Besides, the RF structure design was made using the MWS software Design and Simulation Results In Table 5, the detailed design and simulation results at 5mA for both the EUROTRANS CH DTL and the MAX CH DTL are compared. And in Fig. 10, the first room temperature CH cavities and the first superconducting CH cavities of the EUROTRANS CH DTL and the MAX CH DTL are shown. 16

17 Table 5: CH DTL parameters for EUROTRANS & MAX. EUROTRANS MAX V eff L cell ß avg E a V eff L cell ß avg E a [MV] [m] [MV/m] [MV] [m] [MV/m] RB RT RT RB SC SC SC SC Figure 10: 1 st RT CH cavities (left) and 1 st SC CH cavities (right) for EUROTRANS (top) [20] and MAX (bottom) [21]. 17

18 Figure 11: Beam transport plots of the EUROTRANS (top) and MAX (bottom) CH DTLs. 18

19 In Fig. 11, the maximum transverse beam sizes along the accelerating channel are plotted for EUROTRANS and MAX, respectively. In both cases, a large safety margin is available. At the exit of the MAX injector, the phase spread is ~15 (1.36cm) and the energy spread is ~300keV. In the acceptance plot of the 17MeV, 352MHz Spoke cavity provided by J. L. Biarrotte [22], it can be seen maximally a ~0.6ns (3.44cm) long beam with an energy spread of ±300keV can be accepted Error Studies Error studies have been also performed for the MAX CH DTL. Randomly generated by the LORASR code, the introduced lens and cavity errors are Gaussian distributed and truncated at the 2σ width within the ranges given in Table 6, where QMIS, QROT, VERR, and PERR are indicating transverse lens offset errors, lens rotation errors, tank / gap voltage errors, and tank phase errors, respectively. Table 6: Error settings for the MAX CH DTL. Error Type Error Settings QMIS [mm] ΔX, ΔY =±0.1 QROT [mrad] Δφ x, y =±1.5, φ z =±2.5 VERR [%] ΔU gap =±5, ΔU tank =±1 PERR [ ] ΔΦ tank =±1 The common transverse beam envelopes for 100 non ideal CH DTLs and the additional rms emittance growths caused by the above mentioned errors are plotted in Fig. 12 and Fig. 13, respectively. Not only no beam loss has been observed, but also it s clear that the beam quality at the end of the injector is still kept good. The maximum additional rms emittance growths for the x, y and z planes are only 8%, 12% and 15%, respectively. 19

20 Figure 12: Common transverse beam envelopes with (red) and without (green) errors for the MAX CH DTL. Figure 13: Additional emittance growths caused by errors for the MAX CH DTL. 20

21 5. CONCLUSIONS Fig. 14 shows the schematic layouts of both the 352MHz EUROTRANS injector and the 176MHz MAX injector together (scaled in length). In the RFQ part, the 4 vane structure is now replaced by a shorter 4 rod cavity. In the CH DTL part, now some focusing elements i.e. the 2 nd rebuncher cavity and two solenoids are removed, the triplets are inserted more deeply into the room temperature CH cavities, and steerer and diagnostics are now included. The beam properties at the end of the injector allow transfer into the 352MHz Spoke cavities and further acceleration in the main linac. As a result of all above mentioned new design concepts and changes, the new layout is even 0.8m shorter at half the frequency. Figure 14: The 17MeV injectors for EUROTRANS (top) and MAX (bottom). An overview of the RF power consumption of the main RT cavities for the EUROTRANS and MAX injectors is given in Fig. 15, where the value for the 4 vane RFQ was given by the MWS software with a safety margin of 20%, that for the 4 rod RFQ was estimated using the 21

22 measured shunt impedance of the SARAF RFQ, 67kΩm [16], and those for the RT CHs were obtained from MWS with a safety margin of 15%. Clearly, the total power consumption for the warm part is considerably reduced, and more important that all power losses for the MAX injector are well below 30kW/m, much lower than 50kW/m, a safe value for reliable CW operation proven by the SARAF RFQ [16] EUROTRANS, Copper Power [kw] MAX, Copper Power [kw] EUROTRANS, Copper Power per Length [kw/m] MAX, Copper Power per Length [kw/m] RFQ RT-1 RT-2 Total Figure 15: RF power consumption of the main RT cavities. To sum up, the new injector will have: A safer CW operation o Lower power density (<<50kW/m). o Greatly reduced sparking risk in the RFQ by increasing the minimum gap between electrodes from 2.6mm to 3.6mm and decreasing the inter vane voltage from 65kV to 40kV. o Less components, so less error sources. Still good beam performance o No beam losses. o Small emittance growths. Reduced costs o 4 rod RFQ structure (easy construction, installation, tuning ). 22

23 o Lower copper power (cheaper RF sources & operation, easy cooling). o Less focusing elements. o Less gaps in the cold part (easier and cheaper welding). o Shorter layout. All above mentioned results have shown that the new design concepts and approaches, especially to use 176MHz as the resonant frequency, will lead to a not only cost saving but also more reliable injector for CW operation while keeping the beam dynamics performance satisfying. 6. REFERENCES [1] server.ka.fzk.de/eurotrans/. [2] [3] N. Pichoff, H. Safa, Reliability of Superconducting Cavities in a High Power Proton Linac, Proceedings of the 7 th European Particle Accelerator Conference, Vienna, Austria, pp (June 26 30, 2000). [4] C. Zhang, M. Busch, H. Klein, H. Podlech, U. Ratzinger, J. L. Biarrotte, Reliability and Current Adaptability Studies of a 352MHz, 17MeV, Continuous Wave Injector for an Accelerator Driven System, Phys. Rev. ST AB 13, (2011). [5] C. Zhang, H. Klein, D. Mäder, H. Podlech, U. Ratzinger, A. Schempp, R. Tiede, From EUROTRANS to MAX: New Strategies and Approaches for the Injector Development, IPAC 11, San Sebastian, Spain, pp (September 2011). [6] Dirk Vandeplassche, Jean Luc Biarrotte, Horst Klein, Holger Podlech, The MYRRHA Linear Accelerator, Proceedings of the 10 th International Topical Meeting on Nuclear Applications of Accelerators, Knoxville, USA (April 3 7, 2011). [7] A. Schempp, Habilitationsschrift, Frankfurt University (1990). [8] T. Sieber, PhD Thesis, Frankfurt University (2001). [9] P. Fischer, PhD Thesis, Frankfurt University (2007). [10] U. Ratzinger, Habilitationsschrift, Frankfurt University (1998). [11] H. Podlech, Habilitationsschrift, Frankfurt University (2008). 23

24 [12] C. Zhang, Z.Y. Guo, A. Schempp, R.A. Jameson, J.E. Chen, J.X. Fang, Low Beam Loss Design of a Compact, High Current Deuteron Radio Frequency Quadrupole Accelerator, Phys. Rev. ST AB 7, (2004). [13] C. Zhang, A. Schempp, Beam Dynamics Studies on a 200mA Proton Radio Frequency Quadrupole Accelerator, Nucl. Instrum. Methods Phys. Res., Sect., A, Volume 586, Issue 2, pp (2008). [14] K.R. Crandall, LANL Internal Report, Nr. LA UR (Revised December 7, 2005). [15] L.M. Young, Simulations of the LEDA RFQ 6.7MeV Accelerator, Proceedings of the 1997 Particle Accelerator Conference in Vancouver, B.C., Canada, pp (May 12 16, 1997). [16] Discussions with Dr. A. Bechtold, NTG, Germany. [17] J. L. Biarrotte, MYRRHA LEBT Preliminary Design Report (July 18, 2011). [18] U. Ratzinger, R. Tiede, Status of the HIIF RF Linac Study Based on H Mode Cavities, Nucl. Instrum. Methods Phys. Res., Sect., A, Volume 415, pp (1998). [19] R. Tiede, G. Clemente, H. Podlech, U. Ratzinger, A. Sauer, S. Minaev, LORASR Code Development, Proceedings of the 10 th European Particle Accelerator Conference in Edinburgh, Scotland, United Kingdom, pp (June 26 30, 2006). [20] F. Dziuba, Diplom Thesis, Frankfurt University (2010). [21] Dominik Mäder, Horst Klein, Holger Podlech, Ulrich Ratzinger, Markus Vossberg, Chuan Zhang, Development of CH Cavities for the 17 MeV MYRRHA Injector, Proceedings of the 2 nd International Particle Accelerator Conference in San Sebastian, Spain, pp (May 4 9, 2011). [22] J. L. Biarrotte, private communication. 24