kephart

Plans for a Proton Driver: 

Plans for a Proton Driver Bob Kephart January 29, 2004

Outline: 

Outline Proton Driver Design Studies 8-GeV synchrotron 8-GeV Superconducting Linac  bulk of the talk MI upgrades FLRP: PD WG and preliminary Recommendations Conclusions

Studies of the FNAL Proton Source: 

Studies of the FNAL Proton Source Several studies have had the goal of understanding the limitations of the existing source and suggesting upgrades Proton Driver Design Study I: 16 GeV Synchrotron (TM 2136) Dec 2000 Proton Driver Design Study II: 8 GeV Synchrotron (TM 2169) May 2002 2 MW upgrade to Main Injector May 2002 8 GeV Superconducting Linac: ~Feb 2004 Proton Team Report (D Finley): Oct 2003 Report: http://www.fnal.gov/directorate/program_planning/studies/ProtonReport.pdf Limitations of existing source, upgrades for a few 10’s of $ M. “On the longer term the proton demands of the neutrino program will exceed what reasonable upgrades of the present Booster and Linac can accommodate FNAL needs a plan to replace its aging LINAC & Booster with a new more intense proton source (AKA a Proton Driver, Superbeam, or VIPS= Very Intense Proton Source)

Proton Driver Design Studies: 

Proton Driver Design Studies 8 GeV Synchrotron (TM 2169) Basic plan is to replace the existing Booster with a new large aperture 8 GeV Booster (also cycling at 15 Hz) Takes full advantage of the large aperture of the Main Injector Goal= 5 times # protons/cycle in the MI ( 3 x 1013 1.5 x 1014 ) Reduces the 120 GeV MI cycle time 20% from 1.87 sec to 1.53 sec The plan also includes improvements to the existing linac (new RFQ and 10 MeV tank) and increasing the linac energy (400600 MeV) The increased number of protons and shorter cycle time requires substantial upgrades to the Main Injector RF system Net result = increase the Main Injector beam power at 120 GeV by a factor of 6 (from 0.3 MW to 1.9 MW)

PD: 8 GeV Synchrotron: 

PD: 8 GeV Synchrotron Sited West of the existing booster Twice the shielding of the current booster Large aperture magnets Collimators contain losses to avoid activation of equipment

PD: 8 GeV Synchrotron: 

PD: 8 GeV Synchrotron Synchrotron technology well understood Can be executed quickly May be cheaper than an 8 GeV linac But… Doesn’t replace entire linac  200 MHz PA’s would still be a vulnerability, aging linac equipment still an issue Cycle time is still 15 Hz it would still take 5/15 of a sec to fill MI with 6 booster batches limits upgrades to the MI cycle time (Beam power is proportional to # p/cycle x cycles/sec) Significant interruption of operations to upgrade linac and break into various enclosures (vs Run II)

PD: 8 GeV SC Linac: 

PD: 8 GeV SC Linac Basic concept, design, (& slides) are due to Bill Foster at FNAL Observation: $/ GeV for SCRF has fallen dramatically  can consider a solution in which H- beam is accelerated to 8 GeV in a SC linac and injected directly into the Main Injector Why an SCRF Linac looks attractive: Many components exist (few parts to design vs new booster synchrotron) Copy SNS, RIA, & AccSys Linac up to 1.2 GeV Use “TESLA” Cryo modules from 1.2  8 GeV Probably simpler to operate vs two machines (ie linac + booster) Produces very small emittances vs a synchrotron Delivers high beam powers simultaneously at 8 & 120 GeV Injection into MI is done with 90 turns of small transverse emittance beam (2 p mm-mrad, 95% normalized) which is “phase space painted” into MI (40 p ) aperture in 1 ms MI “fill time” that is negligible vs MI ramp times (more later)

8 GeV Linac Siting for Design Study: 

8 GeV Linac Siting for Design Study Sited tangent to the Main Injector

Multi-Mission 8 GeV Injector Linac: 

Multi-Mission 8 GeV Injector Linac A SC LINAC might also have many other Missions at FNAL eg… accelerate electrons as a 1.5% systems test of a cold Linear Collider

A Draft Design Study exists: 

A Draft Design Study exists Web Link: http://tdserver1.fnal.gov/project/8GeVLinac/DesignStudy/ 131 page document Plan: Next Few Weeks: Merge with PD II Design Study Technically it looks to be feasible Principle issue is the cost SNS was very expensive but there are reasons that this was so… TESLA appears to be very cheap / Gev Need to do a careful Technical Design Report including optimization and costs That’s the plan (more later)

Basic plan for an 8 GeV SC Linac: 

Basic plan for an 8 GeV SC Linac Commercial 402.5 MHz RFQ & DTL up to 87 MeV Accelerator Physics design ~ cloned from SNS 805 MHz Superconducting Linac up to 1.2 GeV Three sections: Beta = 0.47, 0.61, 0.81 Use cavity designs developed for SNS & RIA TESLA-style cryomodules for higher packing factor 1.2 GHz “TESLA” cryomodules from 1.2-8 GeV This section can accelerate electrons as well RF from one Klystron fanned out to 12 cavities Current design study assumed TESLA 500 gradients (25 MV/m) to achieve 8 GeV, if TESLA 800 gradients (35 MV/m) are practical  can operate at 12 GeV or could reduce the cost accordingly

AccSys Source/RFQ/DTL: 

AccSys Source/RFQ/DTL The low RF duty factor of the SC linac means one may be able to buy the linac front end commercially vs design and build it (SNS = expensive) AccSys has shipped multiple RFQ/DTL units for medical purposes in recent years. Front end needed for SC linac is very similar Vendor Estimate is ~$27M base cost for turn-key operation @87MeV. (Less if FNAL provides the RF Power source) AccSys PL-7 RFQ with one DTL tank

Most other TECHNICAL SUBSYSTEM DESIGNS EXIST and have been shown to WORK: 

Most other TECHNICAL SUBSYSTEM DESIGNS EXIST and have been shown to WORK SNS Cavites FNAL/TTF Modulators “TTF Style” Cryomodules Civil Const. Based on FMI RF Distribution* *requires ferrite phase shifter R&D

TESLA-Style Cryomodules for 8 GeV: 

TESLA-Style Cryomodules for 8 GeV Design conceptually similar to TESLA No large cold gas return pipe Cryostat diameter ~ LHC RF Couplers are KEK / SNS design, conductively cooled for 10 Hz operation Cold string length ~ 300m vs every module in SNS => cheaper (more like TESLA)

RF System for 1.2 8 GeV Linac: 

RF System for 1.2 8 GeV Linac Assumes TESLA-style RF distribution works One TESLA multi-beam Klystron per ~12 Cavities Requires a “fast ferrite” E-H tuner to control the phase and amplitude to each cavity The fundamental technology is proven in phased-array radar transmitters. This R&D was started by SNS but dropped due to lack of time. R&D is required to optimize the design for the Linac, funding in TD FY04 budget to start this effort Also needed if Linac alternates between e and P. Modulators are identical to TESLA modulators

RF Fanout at Each Cavity: 

RF Fanout at Each Cavity

ELECTRONICALLY ADJUSTABLEE-H TUNER: 

ELECTRONICALLY ADJUSTABLE E-H TUNER FERRITE LOADED SHORTED STUBS CHANGE ELECTRICAL LENGTH DEPENDING ON DC MAGNETIC BIAS. TWO COILS PROVIDE INDEPENDENT PHASE AND AMPLITUDE CONTROL OF CAVITIES Attractive Price Quote from AFT (<< Klystron)

8 GeV Linac Parameters: 

8 GeV Linac Parameters

Main Injector with 8 GeV Linac: 

Main Injector with 8 GeV Linac H- stripping injection at 8 GeV 25 mA linac beam current 90-turn Injection gives MI Beam Current ~2.3 A ( SNS has 1060 turn injection at 1 GeV ) preserve linac emittances ~2 (or even ~0.5 (95%) at low currents) phase space painting needed at high currents avoids space charge limitations present at lower energy  can put a LOT of beam in MI ! 1.5 Second Cycle time to 120 GeV filling time 1 msec or less no delay for multiple Booster Batches no beam gaps for “Booster Batches” -- only Abort gap Even faster MI cycle times can be considered ( x 2 ?)

120 GeV Main Injector Cycle with 8 GeV Synchrotron: 

120 GeV Main Injector Cycle with 8 GeV Synchrotron

120 GeV Main Injector Cycle with 8 GeV Linac, e- and P: 

120 GeV Main Injector Cycle with 8 GeV Linac, e- and P

Linac Allows Reduced MI Beam Energy without Compromising Beam Power: 

Linac Allows Reduced MI Beam Energy without Compromising Beam Power MI cycles to 40 GeV at 2Hz, retains 2 MW MI beam power

Running at Reduced Proton Energy Produces a Cleaner Neutrino Spectrum: 

Running at Reduced Proton Energy Produces a Cleaner Neutrino Spectrum Running at 40 GeV reduces tail at higher neutrino energies. Same number of events for same beam power  may be a useful operating mode (Plot courtesy Fritz & Debbie)

Fermilab:Long Range Planning: 

Fermilab:Long Range Planning In April of 2003 the Fermilab Director formed a committee to provide advice on the long range scientific program of the laboratory The membership of the LRP committee and its charge can be found at this web site: http://www.fnal.gov/directorate/Longrange/Long_rang_planning.html Excerpt from the charge to the LRP committee: “I would like the Long-range Planning Committee to develop in detail a few realistically achievable options for the Fermilab program in the next decade under each possible outcome for the linear collider. ….“

FLRP:PD Working group: 

FLRP:PD Working group PD Subcommittee: Bob Kephart, chair Steve Geer  Chris Hill                    Peter Meyers Sergei Nagaitsev Technical Advisors Dave Finley Past BD Head (proton economics) John Marriner Past BD Head Shekar Mishra Past deputy head MI project  Victor Yarba SCRF R&D (started TD RF group) Proponents Weiren Chou Synchrotron based Proton Driver Bill Foster SCRF Linac based Proton Driver

FLRP:PD Working group: 

FLRP:PD Working group Had a series of 14 meetings Well attended by Expert Participants 27 additional people made presentations or important contributions to the meetings 3 joint meetings with other LRP sub committees To obtain input from the community an open session took place on Oct 9, 2003 “FLRP Retreat” Jan 9-10 “Preliminary” Proton Driver Recommendations Final Report and recommendations in Feb 2004 PD meetings has now evolved into a regular Proton Driver R&D/Design meeting More people joining the effort

Comparison of PD options: 

Comparison of PD options My conclusions: The SCRF Linac PD is more likely to deliver the desired performance, is more “flexible” machine than the synchrotron based PD, and has more “growth” potential

Preliminary PD Recommendations: 

Preliminary PD Recommendations We recommend that Fermilab prepare a case sufficient to achieve a statement of mission need (CD-0) for a 2 MW proton source (Proton Driver). We envision this project to be a coordinated combination of upgrades to existing machines and new construction. We recommend that Fermilab elaborate the physics case for a Proton Driver and develop the design for a superconducting linear accelerator to replace the existing Linac-Booster system. Fermilab should prepare project management documentation including cost & schedule estimates and a plan for the required R&D. Cost & schedule estimates for Proton Driver based on a new booster synchrotron and new linac should be produced for comparison. A Technical Design Report should be prepared for the chosen technology.

CONCLUSIONS: 

CONCLUSIONS It seems likely that a new intense proton source (AKA Proton Driver) will be proposed for construction at Fermilab in the not too distant future Similar in scope to the Main Injector Project (cost/schedule) An 8 GeV Superconducting Linac appears to be both possible and technically attractive The FNAL management plans to request a complete Technical Design Report for an 8 GeV SC linac including cost & schedule information in the next year This will make it possible to submit a Proton Driver project to the DOE for approval and funding