1. Heavy Ions in RHIC – How to get to higher Luminosities? Angelika Drees APS DNP Fall meeting Oct. 30, 2003 I.Accelerator Basics II.Luminosity Limitations and how to deal with them III.RHIC II Luminosity Upgrade Angelika Drees, 21st Winter Workshop on Nuclear Dynamics, Breckenridge, Feb. 2-12, 2005

2. I. Basics: Luminosity The luminosity can be increased if: o There is more beam/bunch in the two rings (N B,N Y ) o There are more bunches colliding (k b ) o The beam profiles, the size of the beam, at the interaction point, is small (  x,  y ) ->  * Common luminosity detectors (ZDC) are installed in all 4 experiments to measure and monitor luminosity (instantaneous and integrated).

3. Bunched Beam total of 55 bunches per ring 12.8  s per revolution Abort gap Beam is accelerated by Radio Frequency (RF) cavities:  28 MHz for acceleration  200 MHz for storage to reduce bunch length 28 MHz defines the number of “ buckets ” = 360, length is 12.8  s/360=35 ns (or 10 m) Coasting beam: continuous, no bunch structure (debunched), cannot be accelerated Bunched (or captured) beam: every 6 th (3 rd ) bucket, i.e. 55+5 (110+10) bunches per ring with 10 9 Au ions 5 10 9 Cu ions Bunch 1 Bunch 55 “cogging”: lock f rev of both beams and rotate bunches such that they collide at IRs

4. Transition energy crossing RHIC is the first super conducting, slow ramping accelerator to cross transition energy  t (~ 23 GeV): What is “transition” ? ¯below transition fast particles arrive early at the RF ¯with increasing energy fast particles go more and more to the outside (Dispersion!) ¯above transition fast particles arrive late at the RF ¯at transition all particles arrive at the same time: short and unstable bunches! Cross unstable transition energy  t by rapidly changing transition energy (2001) using special quadrupoles: Beam energy RF t t

5. Transition crossing with radial energy jump Dipole oscillations after transition Transition-phase flip

6. II. Luminosity Limitations Luminosity from Beam Parameters: beam/bunch intensity beam sizes Why don’t we just in(de)crease both? The RHIC Cycle and its Challenges: 1.Acceleration: the ramp 2.Store Keep in mind that we cannot make more beam ;) nor (yet) decrease beam sizes once injected and accelerated!

7. Luminosity Limitations Injector performance (routine)/ injections: oAu1  10 9 /bunch, 10  mm mrad oCu 6  10 9 /bunch, 15  mm mrad  additional bunch merge (Booster), damping, new source? Injection/ramp: Vacuum break-downs/electron clouds oAu: < 40  10 9 ions/ring, worse for 110 bunches oCu: with 37x37 pattern and 4.5  10 9 ions/bunch no sign yet  More baking, scrubbing?, NEG coating Ramp: Single bunch instabilities around transition: o Au: < 1.1  10 9 ions/bunch o Cu: so far no sign up to 6  10 9 /bunch  chromaticity control? Store: Intra-Beam Scattering (IBS), beam-beam, debunching oTransverse and longitudinal emittance growth for Au and Cu  stochastic cooling, eventually will need electron cooling Store: Experimental background ocoupling, bump non-closure, halo scraping …  model, 2ndary collimators, gap cleaning RHIC cycle

8. Pressure rise at injection and during the ramp Mainly in warm sections that didn’t have bake-out; worse with 110 bunches/ring Ion desorption, electron desorption, electron multi-pacting, electron cloud Installed electron detectors in IP12 and IP2 and solenoids for electron suppression in IP12. 10 -5 Torr Pressure at IP12 Slow rise. Electron cloud + ? 55 bunches 9x10 8 Au/bunch Pressure at IP12 Run-away rise. Loss induced gas desorption? 2x10 -4 Torr 55 bunches 6x10 8 Au/bunch 50 s 20 s

9. Pressure rise with 100 ns bunch spacing Pressure rise coincides with the electron signal. Solenoid field reduces electron signal. Solenoid field (4 m) partially reduces pressure rise in the 34 m long straight section. Beam intensity [10 11 ] Pressure Rise [10 -8 Torr] Eectron signal [10 mV] Solenoid on Minutes

10. Flexible Bunch Patterns Use a combination of 60-bunch and 120- bunch fill patterns by injecting bunches into 3 rd and 6 th buckets. Total number of bunches is: 55 < 68 < 110 Increase is 20% per ring. Avoid vacuum pressure rise while maintaining same luminosity (increase bunch current!) by using any number of bunches < 56

11. Vacuum: NEG coating NEG strips first used in TTB(BNL) and LEP(CERN) Non-Evaporable Gutter coating: Ti 30 Zr 30 V 40 sputtered ~ 1  m thick onto walls Developed at CERN for LHC warm sections Ultimate pressure < 10 -12 Torr Activation: 1 h @ 250º C, 5 h @ 200º C, 24 h @ 180º C Can be activated numerous times (order of 10 years) Secondary Electron Yield (SEY): 1.1 after activation of 2 h @ 200º C Strong suppression of multi-pacting (tested at SPS) Electron stimulated gas desorption: ~ 100 times lower than baked SS Ion stimulated gas desorption: ~ 10 times lower than SS (tested with 4.2 MeV/n Pb) Test at RHIC: install 60 m of coated pipe, test ion desorption at Tandem NEG strip at TTB

12. Acceleration: Fast Instabilities 01:53:4301:53:44 With Instability Without Instability

13. Emittance Growth during the Ramp beginning of ramp transition end of ramp Two ramps of recent Cu stores bottom: before chromaticity adjustment top: after chromaticity adjustment emittance growth of up to 100% still remains (not understood yet) chromaticity: dq/q=  dp/p (tune change as a function of momentum change)

14. Intra-Beam Scattering (IBS) in RHIC Longitudinal and transverse emittance growth agrees well with model Longitudinal emittance growth causes debunching (bunched beam lifetime) => stochastic cooling will allow up to 50% more beam/bunch Some additional source of transverse emittance growth IBS determines RHIC Au performance Eventually will need electron cooling

15. Beam Growth during a Store blue and yellow emittances during 3 Cu stores, Jan. 30, 2005 Norm. emittances between 15 and 20  mm mrad grow up to 35  mm mrad at the end of store  => IBS + beam-beam bunch by bunch emittance measured during a Cu store Jan 31, 2005. Bunch1 is affected by the gap cleaning procedure, bunch 19,37,121 have 4 collision points, bunch 100 has only 3 => beam- beam causes additional emittance growth.

16. Gap Cleaning cleaning starts blue total beam blue bunched beam 30 min. was: gap cleaned at the end of store (> 30 min.!), bunch by bunch in gap now: continuous excitation (1 Hz) through store, kicking ~150 ns of gap. “Gap Cleaning” means minimization of the debunched beam (I total -I bunched ) to stay within operational envelope (uncontrolled beam loss/hour) and to prevent magnet quenches at time of beam dump. Abort Gap Cleaning in 2 steps: excite the debunched beam (in gap) transversely with the tune meter kickers @ IR2 remove the excited beam with the transverse collimators @ IR8

17. Cu-Stores with and w/o continuous gap cleaning cleaning in middle and at end of store => loss of time for integrated luminosity continuous cleaning allows physics w/o interruption => risk of higher backgrounds cleaning

18. Cu FY05 run integrated luminosity so far … maximumminimum  *(m) 1.0 0.9 2.5 2.8 one week ago: yesterday delivered/accepted varies for experiments!

19. RHIC HI– achieved parameters Mode No of bunches Ions/bunch [  9 ]  * [m] Emittance [  m] L peak [cm -2 s -1 ] L store ave [cm -2 s -1 ] L week Au-Au [Run-4] 451.1110 15  10 26 5  10 26 160  b -1 d-Au [Run-3] 55110/0.7115 12  10 28 3  10 28 4.5 nb -1 Cu Cu [Run-5] 3750.920-25few 10 28 n.a. Au-Au design561215-409  10 26 2  10 26 50  b -1 [best store or week] [incl. beam experiments and maintenance] Enhanced HI Luminosity Goal (before e-cooling, about 2008, assume stochastic cooling) For Au-Au, average per store, 4 IRs: L = 8  10 26 cm -2 s -1 at 100GeV/u 4x design 2  achieved

20. Projected 4-year Au-Au luminosity Assume - 12 weeks production in every year - 8 weeks of linear luminosity increase - 4 experiments - completion of improvements 2  increase

21. III. RHIC II luminosity upgrade RHIC luminosity is limited by intra-beam scattering  beam cooling at full energy! Needs improved injector -> EBIS Feasibility study by BINP (V. Parkhomchuk et al.): RHIC luminosity can be increased ten times. Bunched electron beam requirements for 100 GeV/u gold beams: E = 54 MeV, ~ 100 mA, electron beam power: ~ 5 MW! Requires high brightness, high power, energy recovering superconducting linac First linac based, bunched electron beam cooling system used at a collider First high p t electron cooler to avoid recombination of e - and Au 79+ Maintains present bunch spacing (~ 100 ns) and available IR length Increased luminosity for pp and other species Longitudinal cooling possibly gives shorter diamond length

22. EBIS/Linac RHIC Pre-Injector Highly successful development of Electron Beam Ion Source (EBIS) at BNL EBIS allows for a reliable, low maintenance Linac-based pre-injector replacing the Tandem Van de Graaffs Produces beams of all ion species including Uranium and polarized He 3 (for eRHIC) Ready to start construction; Cost: 16.1 M$; Schedule: 2005/6 – 2008/9 EBIS test stand

23. EBIS layout

24. RHIC II Luminosities with e-Cooling LinacRf Gun Buncher Cavity Cooling Solenoid (~ 30 m, ~ 1 T) Debuncher Cavity e-Beam Dump Gold beam Gold collisions (100 GeV/n x 100 GeV/n): w/o e-coolingwith e-cooling Emittance (95%)  m15  40 15  3 Beta function at IR [m]1.01.0  0.5 Number of bunches112112 Bunch population [10 9 ]11  0.3 Beam-beam parameter per IR0.00160.004 Peak luminosity [10 26 cm -2 s -1 ]3290 Average luminosity [10 26 cm -2 s -1 ]870 demonstrated by JLab for IR FEL. (50 MeV, 5 mA)

25. RHIC Luminosity with and without Cooling Transverse beam profile during store Leveling of luminosity and beam-beam interaction through continuous cooling and beta squeeze 2 mm 5 hours With e-cooling Without e-cooling 0.0 0.5 1.0 1.5 2.0 2.5 Luminosity [10 28 ]

26. Summary  Successful operation of RHIC with 100 GeV/n HI beams in three modes (at various energies): Gold – gold collisions, peak luminosity = 5  10 26 cm -2 s -1 d–Au collisions, peak luminosity = 7  10 28 cm -2 s -1 Cu-Cu collisions,peak luminosity = ?  10 28 cm -2 s -1  Full exploitation of the many physics capabilities of RHIC requires extensive, ongoing machine development  Accelerator Improvement Projects are required to improve performance and maintain and enhance reliability of the RHIC operation  Successful EBIS R&D allows for the construction of a new Linac-based RHIC pre-injector  RHIC e-cooling R&D program started for 40 times RHIC Au-Au design luminosity upgrade  Luminosity can be increased by another factor x10 => will you be ready to take it?

27. Supplemental

28. RHIC Run-5 Run-5 (started in November 2005)  Cu – Cu at 100 GeV/u – 4 weeks set-up / ramp-up – 7 weeks (?) for luminosity – 7 nb -1 integrated luminosity delivered to Phenix/Star (assuming same charge per beam as with Au in Run-4)  p  - p  at 100 GeV/u – 3 weeks set-up / ramp-up – 8 weeks (?) for luminosity – 8 pb -1 integrated luminosity delivered to Phenix/Star, with 45% polarization at store (demonstrated performance of Run-4) [T. Roser, W. Fischer, M. Bai, F. Pilat, “RHIC Collider Projections”, Last update August 16, 2005.]

29. Low Energy Running

30. Scraper and Kicker Location in RHIC

31. The Zero Degree Calorimeter •3 modules on either side of the Interaction Region (IR) •Same detector at all IRs with experiments. •About +/- 18 m distance from center behind DX => Only neutron sensitive •Covers +/- 2.5 mrad forward angle

32. Storage: Measuring Luminosity Zero-Degree Calorimeters Protons bend away Detects energy from neutrons Timing is Everything

33. RHIC upgrades for Enhanced Luminosity – risk assessment SpeciesLimitFurther mitigationRiskAlternatives Au, pWarm vacuum1. NEG coated pipes 2. Other (1) Low (2) Solenoids, ion pumps pCold vacuum1. Pre-pumping 2. Beam-scrubbing (3) MediumNo alternatives pBeam-beam (4) 1. Orbit feed back for IP (5) 2. Better  Q min,  control Medium?? Notes: (1) Any workable solution at bottlenecks (solenoids, other coating, redesign for baking, …). (2) Successive layers of bottlenecks may not be removed fast enough (one layer per shut-down). (3) Beam scrubbing for SEY reduction demonstrated by CERN. If e-clouds release chemisorbed molecules they can only be pumped out after warm up. This may need too many thermal cycles (at most one per shut-down). (4) Need to accommodate  Q=0.015 with luminosity lifetime > 10 hours. Selected new working point, concluded that change of phase advance between IPs will not result in any gain. (5) To suppress 10Hz oscillation of about 5% of rms beam size, suspected to lead to emittance growth.

34. Electron Cooler Beam Dynamics R&D Use two solenoids with opposing fields to eliminate coupling in the ion beam. A quadrupole matching section between the solenoids maintains magnetization. Stretcher / compressor with large M56 and zero M51, M52 Merge beams with two weak dipoles with Stabenov solenoid focusing to minimize dispersion and avoid coupling.

35. 1. New Collimation System set of 4 PinDiodes White Secondary Collimators are Vertical Steering and motion of collimators (total of 18 motion channels!) will be automated by feedback based on signals from PinDiodes (loss monitors). Status: collimators ready to be installed, software development in progress.

36. Pressure rise with 100 ns bunch spacing Pressure rise coincides with the electron signal. Solenoid field reduces electron signal. Solenoid field (4 m) partially reduces pressure rise in the 34 m long straight section. Beam intensity [10 11 ] Pressure Rise [10 -8 Torr] Eectron signal [10 mV] Solenoid on Minutes