Front-End System: Ion Source and RFQ Accelerator Derun Li Center for Beam Physics Lawrence Berkeley National Laboratory September 11-12, 2009 Fermi National Accelerator Laboratory

Acknowledgments Special thanks to John Staples, Qing Ji, Steve Virostek, John Byrd, John Corlett, Steve Gourlay and colleagues at Center for Beam Physics of LBNL for their contributions and suggestions to this presentation and Sergei Nagaitsev and colleagues at FNAL for collaboration studies of the front-end system September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Front-End System of Project X Ion source (H- CW 10 mA) LEBT RFQ 2.5 MeV MEBT September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Contents LBNL’s ion source experience and the state-of-art ion sources Proton sources Negative ion sources Proposal of CW H- ion source for Project X RFQ accelerator experience & history Five RFQs have been built so far Recent RFQ research A 6 MeV deuteron RFQ for neutron source (as an example) Innovative beam dynamics design Improved RF structure design 3 D E&M simulations Mechanical design and fabrication CW RFQ studies for Project X (John Staples’ presentation) Summary We propose to design, construct, deliver and commission complete front- end system for Project X September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

LBNL’s Ion Source Experience Extensive experience and highly successful history in ion source and injector development. Past examples include the positive and negative ion sources for neutral beam injectors for Fusion research programs, and the H- ion source for SSC and SNS. Proton sources Extensive experience in RF-driven proton source (in both CW and pulsed mode) development Two types of H- ion sources developed at LBNL: Surface-conversion type H- sources (LANSCE, KEK ….) Volume-production type H- sources (SSC, SNS ….) September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Proton Source Mature technology for producing 10s mA CW proton beams. We propose to use RF-driven Multicusp Proton Source Permanent Magnets Antenna Cylindrical chamber wall Quartz RF window Back streaming electron dump 10 cm September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Surface Conversion H- Ion Source Employed in LANSCE Upgrade* 1 ms pulse at 120 Hz Extracted peak current versus Arc power * M. Williams et al, “Ion Source Development for LANSCE Upgrade”, 19th International Linear Accelerator Conference, Chicago, IL, USA, August 23-28, 1998. September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Status of LANSCE Ion Source Current LANSCE H- ion source (operational)* Two filaments (250 mm long, 1.6 mm thick, high purity tungsten) 4-11 kW discharge power Beam current 14-24 mA through a 9.8 mm aperture Emittance < 0.2 π mm-mrad (95%, normalized, 1-rms) Duty factor 6-12 % Current R&D on RF (13.56 MHz) helicon-driven converter source at LANL Beam current 7-12.5 mA through a 9.8 mm aperture Duty factor 1-3.7% demonstrated * R. Keller et al, “H- ion source development for the LANSCE Accelerator Systems”, 1st International Symposium on Negative Ions, beams, and Sources, Aix-en-Provence, France, September 9-12, 2008. September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Volume-production H- Ion Source RF-driven H- Source with Porcelain Antenna (currently been used in the SNS at Oakridge National Laboratory) September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

SNS Electrostatic LEBT The ion source with outlet and dumping electrode is tilted by 3 degree with respect to the LEBT axis. Some magnet orientations are rotated into the viewing plane of this illustration. The shape of the beam envelope is exaggerated for emphasis. September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Status of SNS H- Ion Source Current SNS H- ion source* RF-driven with internal antenna Routinely delivering 32 mA 0.67 ms, 60Hz, ~ 4% duty-factor Development of RF-driven external antenna ion source at SNS Demonstrated maximum beam current of 95 mA at 52 kW RF power, with Ni elemental collar 1 ms pulse at 60 Hz * R. F. Welton et al, “Next Generation H- Ion Sources for the SNS”, 1st International Symposium on Negative Ions, beams, and Sources, Aix-en-Provence, France, September 9-12, 2008. September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Status of CW H- Sources Operational CW H- ion source in accelerators Multicusp filament TRIUMF 1-5mA (0.5 π mm mrad) Jyvaskyla-LIISA: 3 mA (0.6 π mm mrad) D-Pace : 15-mA, 30 kV DC H- (< 1 π mm mrad, 4 rms) The state-of-art CW H- ion source* Reported by Yu. Belchenko et al Technology: Penning discharge, surface-plasma source Current: 15 mA Normalized 1-rms emittance < 0.2 π mm mrad * Yu. Belchenko et al, “A 15 mA CW H- source for accelerators”, 1st International Symposium on Negative Ions, beams, and Sources, Aix-en-Provence, France, September 9-12, 2008. September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

RF-driven Volume H- Ion Source with External Antenna (pulsed) H- Sources for Project X RF-driven Volume H- Ion Source with External Antenna (pulsed) RF-driven Surface Conversion H- Ion Source (CW) Cesium collar Cesium oven and injector Antenna Cylindrical Quartz wall Converter Repeller H- beam Cylindrical chamber wall Antenna Quartz RF window Filter magnets Permanent Magnets September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

CW H- Sources for Project X We propose to use the surface conversion ion sources for CW operation: Switching from filament discharge to RF-driven Long lifetime Cleanliness Emittance optimization Flat converter or concave converter Less extracted electron beam Low risk for CW operation September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

LBNL has designed and built five RFQs LBNL RFQ Experience & History LBNL has designed and built five RFQs RFQ1 - Bevatron RFQ2 – CERN RFQ3 – BNL Injector Accelerator Handbook RFQ Article US Particle Accelerator School RFQ4 – LBNL Proton RFQ5 - SNS September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Most Recent RFQ Research A 6 MeV Deuteron RFQ for Cargo-Screening Evaluating two beam dynamics approaches (John Staples’s presentation) Modified “conventional” design (LANL method) Like SNS RFQ, but significantly shortened Kick-buncher+drift design (LBNL method) IUCF RFQ for cyclotron injector based on this design Length in the 5 meter range, 50-60 kV injector voltage Similar in beam characteristics, each will satisfy requirements Fabrication techniques slightly different We have developed many mechanical concepts over the 28 year history of LBNL RFQ development Best ideas from the last two RFQs and combine them for this and future designs September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Two Representative RFQ Models RFQ 4 (BNL) – 750 keV proton Bolt-together construction Copper-plated aluminum Vane coupling rings ~ 1 meter long RFQ 5 (SNS) – 2.5 MeV proton 6% duty factor Cu/Glidcop brazeup Pi-mode stabilizers 4 segments, 3.7 m long September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Cavity Construction The SNS segmented structure with pi-mode stabilizers works well, but is expensive to manufacture Developed for very long life, high duty factor The 750 keV proton structure with vane coupling rings is efficient and compact Developed for low-cost manufacturing techniques Take the best aspects of each and combine them into a structure with good thermal characteristics, good field stability and good thermal properties, as well as easy to fabricate September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

RFQ Beam Dynamics Designs We use three independent codes to design Parmteq-h (LBNL) Toutatis (CEA-Saclay) Parmteq-M (LANL) The innovative kick-buncher to beam dynamics design has been thoroughly checked with independent codes Beam dynamics extensions include Dynamic tapering of parameters along z Stepping physical parameters in each module for ease of construction September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Parmteq Beam Dynamics Simulation - RFQ length: 5.05 m - Tracking of 5000 particles - Total # of cells: 225 - The RFQ parameters (focusing and m) are adjusted to give the best transmission and exit beam parameters - Low peak surface field - Evolution of the transverse beam sizes - Phase and energy spread along the structure September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Runs slowly, but has more detailed physics model Toutatis Beam Dynamics Calculation Runs slowly, but has more detailed physics model September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Better E&M Modelling Tools Now More powerful 3 D Electromagnetic codes available now Extensive experience in CST Microwave Studio Much better mesh with adaptive techniques Benchmarked the code against the SNS RFQ* Excellent agreement, even with complex end sections Good enough to skip cold modelling stage Still need to do a final tune with bead pull RFQTUNE program facilitates setting tuners Get frequency within klystron power bandwidth * D. Li, J. Staples and S. Virostek, DETAILED MODELING OF THE SNS RFQ STRUCTURE WITH CST MICROWAVE STUDIO, Linac-2006, Knoxville, TN September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

SNS RFQ with Microwave Studio Frequency to 0.5% Field profile correct Complex end tuner geometry Pi-mode stabilizers September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

RFQ Structure Simulations Parameterized 3D CST MW model RFQ with -mode stabilizers A 3D RFQ slice for quick frequency search, cross-section optimization and RF parameter calculations A 3D 1.28 m long RFQ model to study -mode stabilization and mode separation Simulations with radial matcher, cut-backs and ends September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

RFQ Structure Simulations Frequency domain simulations Field distributions and mode separations Mode separation ~ 20 MHz Quadrupole mode f = 200.103 MHz Dipole mode fH = 224.359 MHz September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

RFQ Detailed 3D CAD Model Sensing loop -mode rods Slug tuners Module joint bolt pockets Minor vane Vane-vane RF seal 3D O-ring Module-module RF seal Major vane Vacuum port September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

3D CAD Model Details Detailed geometry of vanes for all four modules and the associated components are included in the model Modulation information is included in a separate data file to generate the input program to the computer controlled milling machine Layout developed to include even spacing of pi-mode rods, tuners and sensing loops along entire RFQ Other design details to be incorporated in the model are: final vane cutback geometry, vacuum port grill and vane- vane bolting, … September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

RFQ fabrication (BNL RFQ) Assembly Cavity and vane fabrication Bead-pull measurement September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

CW RFQ Studies for Project X More details in John Staples’s presentation Overview of CW RFQs Frequency choices Beam dynamics designs Comparison between “conventional” and innovative “kick-buncher” designs CW RF cavity/structure and mode stabilization Thermal management Mechanical stability Minimize required RF power MEBT considerations … We do collaboration better than anyone September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory

Summary LBNL has extensive experience in ion source development and RFQ design and fabrications Expertise in beam dynamics, engineering, fabrication and commissioning We are willing to collaborate with FNAL on Project X and propose to design, construct, deliver and commission whole front-end system Ion source (CW H-) and LEBT RFQ accelerator (CW) and MEBT Ancillary components to produce a fully integrated system Experience in multi-lab collaborations Started detailed RFQ studies for Project X front-end We do collaboration better than anyone September 11-12, 2009 Front-End System: D. Li, Lawrence Berkeley National Laboratory