Snowmass G. Eigen, University of Bergen G. Eigen University of Bergen Snowmass July 14, 2001

Snowmass G. Eigen, University of Bergen OUTLINE  Introduction  Silicon Vertex Detectors  Drift Chambers  DIRC  Electromagnetic Calorimeters  IFR  Trigger Rates  Examples of strawman detectors  Conclusion

Snowmass G. Eigen, University of Bergen Introduction  For precision measurements of CP-violation asymmetries and rare B decays high luminosities are an important prerequisite  Recently, the design of an e + e - storage ring  s ~ 10GeV with luminosities of £ peak = has become feasable  So how do present subsystems of multipurpose detectors cope with the increased background levels ?  The following results are my personal views based on BABAR studies: Report of the High-Luminosity Background Task force (C. Hast, W. Kozanecki (chair), A. Kulikov, T.I. Meyer, S. Petrak, T. Schietinger, S. Robertson,M. Sullivan, J. Va’vra, BaBar Note 522)  Results for £ peak ≥ should be taken with a grain of salt:  Extrapolations are made over > 2 orders of magnitude (errors > factor 2)  Extrapolations depend very much on IR layout

Snowmass G. Eigen, University of Bergen Background Issues  Acceptable levels of backgrounds are determined by  Radiation hardness of subdetectors  inefficiencies, destruction  Trigger rate  deadtime, loss of signal  Detector occupancies  inefficiencies, worse resolution, worse S/B  Occupancy and trigger rate determine acceptable dynamic running conditions  Total integrated radiation dose determines lifetime of subdetectors  Dose is accumulated under normal running conditions, during injection, machine studies and beam-loss events  At PEP II dose accumulated during running dominates  At £ peak = machine this is different, injection losses determine dose

Present Measures of BABAR Subsystems to Machine Backgrounds  Radiation Hardness of SVT detector modules is estimated at 2MRad Instantaneous dose rate in radiation protection diodes BW:MID & FE:MID are within factor of two representative of harshest radiation levels hitting SVT modules in horizontal plane  Total current drawn by drift chamber is limited to 1000  A by existing HV power supplies  Counting rate above kHz in DIRC phototubes starts inducing significant dead time with present electronics  Fractional EMC crystal occupancy above a 1 MeV threshold and number of crystals above 10 MeV characterize potential degradation of calorimeter energy resolution, as well as number of fake neutral clusters  Level-1 (L1) trigger rate is currently limited to kHz by DAQ bandwidth considerations

Snowmass G. Eigen, University of Bergen Present Sources of Machine Backgrounds  Detector subsystems are subjected to different machine-related backgrounds  Electrons:  lost particles backgrounds (beam-gas bremsstrahlung,  Coulomb scattering) and synchrotron radiation  Positrons:  lost particles backgrounds (beam-gas bremsstrahlung)  2 beams:  no collision  single beam backgrounds above plus beam-gas cross term  in collision  backgrounds from luminosity, beam-beam tails & above 3  Note that there is a difference in operation between PEP II at high £ peak & an  £ peak = collider:  PEP II: inject & run (stable beams)   continuous injection (no stable beams)

Snowmass G. Eigen, University of Bergen Backgrounds in PEP II & in an L=10 36 Machine  Background estimates by W. Kozanecki based on J. Seeman’s design PEP II  Beam loss rates in PEP II and a machine differ by a factor of > 10 3 but only small fraction will contribute to detector backgrounds HER LER Super HERSuper LER Beam current I b [A] Beam lifetime  b [min] Beam loss rate I b /  b Luminosity [A/min] Vacuum [A/min] Touschek [A/min] b-b tune shift [A/min] Dynamic aperture [A/min] Total [A/min] 1.3  

Snowmass G. Eigen, University of Bergen Backgrounds in an L=10 36 Machine  In PEP II LER lifetime is dominated by vacuum or Touschek effect, while HER lifetime is affected by beam-beam tune shift and then vacuum  Background sources in SVT, DCH and EMC result from beam-gas in the incoming straight section  Beam-beam tune shifts, dynamic aperture and vacuum losses probably will contribute to vacuum-like backgrounds, since losses are transverse (like distant LER Coulomb scattering in PEP II)  Since quads need to be shielded transverse losses are produced at betatron collimators far from IP  combined transverse losses are main issue with LER backgrounds only 15%-20% of them, HER is minor problem  Effects of longitudinal losses at £ peak = are not known, since these have not been studied in PEP II  Since sum of longitudinal, all transverse and injection losses is so large, vacuum in IR will be less a problem, still need pressure of within 50m of IP

Snowmass G. Eigen, University of Bergen Dependence of backgrounds on beam currents  Touschek: Dependence is not known, expect no effect for a while (low I LER, long Touschek lifetime, negligible secondary particles) At some point it will take off  need simulation with Turtle  Beam-beam tune shift: very non-linear and very tune sensitive  Dynamic Aperture: linear (?)  Vacuum: quadratic in I LER (the base pressure will be well-controlled, the dynamic pressure will dominate)  Luminosity: linear in beam currents

Snowmass G. Eigen, University of Bergen Estimates of backgrounds due to beam losses  Touschek: Need Turtle-like simulation of energy spectrum  Tranverse losses: Scale distant Coulomb prediction by the ratio of loss rates with measured distant LER-only contributions (DCH,DIRC)  Injection losses: Take clean injection day from PEP II and scale by injection currents  Secondary particles: Due to multistage injection, betatron collimation and momentum collimation secondary particles are big issue realistic simulation is a major task  Radiative Bhabha: Debris in the detector from radiative Bhabhas eventually will become large, it is sensitive to beam line geometry & IR layout  Caution: In extrapolations below none of above effects is included

Multipurpose Detector for e+e-Collisions at 10GeV

Snowmass G. Eigen, University of Bergen Luminosity Considerations  For luminosities shown in blue extrapolations have been taken from the report of the High-Luminosity Background Task force, while for luminosities shown in green results are my extrapolations using the algorithms given by the High-Luminosity Background Task force Date£ peak [cm -1 s -1 ] June  August  ? 5.0  ? 1.0   10 36

Snowmass G. Eigen, University of Bergen Silicon Vertex Trackers

Snowmass G. Eigen, University of Bergen Dose accumulated in BABAR SVT

Snowmass G. Eigen, University of Bergen SVT Radiation Dose in Middle Plane time SVT dose rate [krad/y] 2  FE MID BW MID  SVT dose rate: FE MID [kRad/y] =128 I LER + 16 I 2 LER BW MID [kRad/y] =246 I HER I 2 HER  In top & bottom planes dose rate is ~ factor of 10 lower than in middle plane

Snowmass G. Eigen, University of Bergen Silicon Vertex Detector Occupancy

Snowmass G. Eigen, University of Bergen Conclusion on Silicon Vertex Detectors  Radiation levels depend very strongly on IR layout, (KEKB < PEP II)  In BABAR silicon detectors are expected to survive a total dose of 2MRad  With replacements of detectors in the MID plane BABAR SVT is expected to survive luminosities of   LHC R&D demonstrated that Si detectors can survive high irradiation H. Yamamoto bonded 150  thick pixels (55   55  ) (CMOS)  At £ peak ~ 1  occupancy is an issue for Si strip detectors close to IR  pixels in first two layers  So for £ peak ~ 1-10  appropriate silicon detectors probably work £ peak [cm -1 s -1 ] 6.5      ∫£dt [fb -1 ]/y I LER/ I HER [A]2.8/1.13.7/1.34.6/1.59/2.518/5.5 D SVT [kRad/y]480/280690/ / / /1630

Snowmass G. Eigen, University of Bergen Drift Chambers

Snowmass G. Eigen, University of Bergen  Machine backgrounds affect operation of Drift Chamber in 3 ways:  Total current I DCH in Drift Chamber drawn by wires is dominated by charge of beam-related showers I DCH is limited by high-voltage system,  above limit chamber becomes non operational!  high currents also contribute to aging of chamber!  maximum Q max: Cb/cm of wire  Occupancy in Drift Chamber due to backgrounds (hits, tracks) can hamper reconstruction of physics events  Ionization radiation can permanently damage read-out electronics & digitizing electronics Drift Chambers

Snowmass G. Eigen, University of Bergen Drift Chamber Currents  I DCH [  A] = 35.3 I LER I 2 LER I HER I 2 HER £ -14 with currents in [A] and luminosity in units of [10 33 cm -1 s -1 ]  Single beam and collision measurements taken June/ July at HV=1900V  For HV=1960V scale current by factor 1.67

Snowmass G. Eigen, University of Bergen Measured Drift Chamber Currents & Models  Single-beam measurements (LER) taken with BABAR DCH in June and July 2000 at HV=1900V

Snowmass G. Eigen, University of Bergen Drift Chamber Backgrounds  Extrapolation for HV=1900V  At HV=1960 background levels are expecetd to be 65% higher time total DCH current [  A] 2  Luminosity I LER I HER

Snowmass G. Eigen, University of Bergen Drift Chamber Occupancy  N DCH = I LER I 2 LER I HER I 2 HER £ with occupancy in [%], currents in [A], luminosity in units of [10 33 cm -1 s -1 ] at 1900V  At HV=1900V (Jan-July): N DCH = I DCH (<350  A)  At HV=1960 V(July-now): N DCH = I DCH (>200  A)  Large spread  extrapolation difficult data points at ~ same £

Snowmass G. Eigen, University of Bergen Drift Chamber Occupancy  Extrapolation for HV=1900V time DCHoccupancyt [%] 2  Luminosity I LER I HER  N DCH = I LER I 2 LER I HER I 2 HER £

Snowmass G. Eigen, University of Bergen Conclusion on Drift Chambers  Total dose depends on ∫£dt: at 20 fb -1 accumulated 100 rads  For £ peak > 1  it is very unlikely that drift chambers will work One needs other devices: straws, TPC with GEM readout, Si tracker £ peak [cm -1 s -1 ] 6.5      ∫£dt [fb -1 ]/y I LER/ I HER [A]2.8/1.13.7/1.34.6/1.59/2.518/5.5 I DCH [  A] N DCH [%] Q wire [mCb]~

Snowmass G. Eigen, University of Bergen GEM Layout

Snowmass G. Eigen, University of Bergen GEM Layout

Snowmass G. Eigen, University of Bergen DIRC PARTICLE IDENTIFICATION

Snowmass G. Eigen, University of Bergen Composition of DIRC Background time DIRC occupancy [ kHz] 2   N DIRC [kHz] = 35 I LER I HER + 25 £ total I LER I HER

Snowmass G. Eigen, University of Bergen Conclusion on DIRC  BABAR DIRC is ok up to £ peak =6  10 34, however the water tank provides a huge Cherenkov detector  At high luminosities £ peak >1  another approach is needed: a compact readout using focussing or timing £ peak [cm -1 s -1 ] 6.5      ∫£dt [fb -1 ]/y I LER/ I HER [A]2.8/1.13.7/1.34.6/1.59/2.518/5.5 N DIRC [kHz]

Snowmass G. Eigen, University of Bergen Different DIRC Imaging Methods  Note that different imaging methods can be chosen in each space dimension

Snowmass G. Eigen, University of Bergen DIRC Readout  B. Ratcliff

Snowmass G. Eigen, University of Bergen Separation Performance vs Random Rates

Snowmass G. Eigen, University of Bergen ELECTROMAGNETIC CALORIMETER

Snowmass G. Eigen, University of Bergen Average Occupancy in EMC Crystals  N EMC (E> 1MeV)= I HER +2.2 I LER £ N EMC (E> 10MeV)= 4.7 I HER I 2 HER +2.4 I LER I 2 LER £ with beam currents in units of [A] and luminosity in units of [10 33 cm -1 s -1 ] Single Crystal occupancy # Crystals with > 10 MeV

Snowmass G. Eigen, University of Bergen Light Yield Changes in EMC

Snowmass G. Eigen, University of Bergen Worst Dose Rate in EMC

Snowmass G. Eigen, University of Bergen Effect of Background on  0 Reconstruction  Background photons both increase  0 background levels and degrade mass resolution

Snowmass G. Eigen, University of Bergen time 2  Composition of EMC Backgrounds EMC occupancyt [%] time # EMC crystals  > 10 MeV total I LER I HER 2   > 1 MeV  N EMC (E> 1MeV)= I HER +2.2 I LER £ N EMC (E> 10MeV)= 4.7 I HER I 2 HER +2.4 I LER I 2 LER £ total I LER I HER noise

Conclusion on Electromagnetic Calorimeters  For luminosities < 1.5  integrated radiation dose for CsI(Tl) crystals is not expected to be a problem if observed light losses scale as expected  Impact of large number of low-energy photons on EMC energy resolution depends on clustering algorithm, digital filtering, etc (needs further study) Expect luminosity contribution to be dominant  Expect reduction of background rates through improvements of vacuum near IR combined with effective collimation against e+ from distant Coulomb scattering  For luminosities >1  light loss due to radiation and occupancy levels for present CsI(Tl) crystals are not acceptable  need R&D studies and look into other scintillator (pure CsI, LSO, GSO?) £ peak [cm -1 s -1 ] 6.5      ∫£dt [fb -1 ]/y I LER/ I HER [A]2.8/1.13.7/1.34.6/1.59/2.518/5.5 N EMC [%] N cluster

Snowmass G. Eigen, University of Bergen Properties of Scintillating Crystals

Snowmass G. Eigen, University of Bergen INSTRUMENTED FLUX RETURN

Snowmass G. Eigen, University of Bergen Conclusion on IFR  Main issue is high occupancy in outer layers due to beam-related backgrounds  Presently outer RPC layer has random occupancy of several %  At design currents and at higher luminosity this will become an unacceptably high contribution to  /  misidentification  Solution for £ peak ~ 3-5  : build 5 cm thick Fe shield following outer-most chamber  At £ peak > 1  occupancy becomes an issue despite shielding RPC’s are not suited, replace them with scintillating fibers

Snowmass G. Eigen, University of Bergen TRIGGERS

Snowmass G. Eigen, University of Bergen L1 Trigger Rate vs Current in Machine

Snowmass G. Eigen, University of Bergen Trigger Rates time Total trigger rate [Hz] 2  Total (L) I HER I LER Background  Expected L1 trigger rate: L1 [Hz]=130 (cosmics)+ 130 I LER I HER + 70 £

Snowmass G. Eigen, University of Bergen Extrapolation on Trigger Rates  For £ peak ~1.5  in BABAR trigger needs to be upgraded to cope with high rates  For higher luminosities one could do more stringent prescaling of Bhabhas, radiative Bhabhas, beam gas, (want to keep all b, c decays) One needs to design appropriate tracking device used in trigger  LHC experiments can accept L1 trigger rates of 100 kHz (ATLAS) bunch crossing is 40 MHz £ peak [cm -1 s -1 ] 6.5      ∫£dt [fb -1 ]/y I LER/ I HER [A]2.8/1.13.7/1.34.6/1.59/2.518/5.5 L1 [Hz]

Snowmass G. Eigen, University of Bergen Trigger for High Luminosity Machine

Snowmass G. Eigen, University of Bergen Trigger for High Luminosity Machine

Detector Considerations  The angular acceptance is limited by beam focussing elements to 300mr  By keeping present boost  =0.58 and a resolution improved by a factor of two one needs to move closer to IP  1cm gold-plated Be beam pipe  To cope with occupancy problems near IR, use Si pixel detectors for first 2 layers of vertex detector, 3 layers Si strip detectors  For central tracker consider either all Si strips, straw tubes or TPC with GEMs readout  For particle identification consider Super DIRC  For EMC consider scintillating crystal calorimeter based on pure CsI, LSO or GSO  For IFR use Fe plates read out with scintillating fibers  Strawman designs resulted in discussions in breakout sessions: G. Dubois-Felsman, G. E., M. Giorgio, D. Hitlin, X. Lou, D. Leith, E. Paoloni, I. Peruzzi, M. Piccolo, M. Sokoloff, H. Yamamoto

Snowmass G. Eigen, University of Bergen Modified Multipurpose Detector TPC with GEMs ECs or strawtubes Pure CsI with APD’s readout Compact DIRC IFR with scintillating fibers First 2 layers pixels +3 layers Si strips

Snowmass G. Eigen, University of Bergen Compact Multipurpose Detector SVT 2 Ly pixel 3 Ly Si strip 4 Ly Si strip tracker Compact DIRC LSO EMC 3T Coil IFR Fe + scint. fibers

Snowmass G. Eigen, University of Bergen Conclusions  Vertex detectors:  Based on studies at LHC silicon vertex detectors probably will work at high luminosties of £ peak ~ 1-10  10 35, need pixel detectors in first two layers (  R & D)  Central tracker:  For £ peak > 1  it is very unlikely that drift chambers will work Need to consider an all Si strip tracker, straw tubes or TPC/GEMs  Particle ID:  With appropriate design of accepted counting rates, beam collimation & shielding a compact DIRC probably will work at £ peak ~ 1-10   Electromagnetic Calorimeter:  For £ peak >1  light loss due to radiation and occupancy levels for present CsI(Tl) crystals are not suitable explore other scintillators (pure CsI, LSO, GSO,…) (  need R&D)  Trigger:  It should be possible to design trigger system for £ peak = 1  10 36

Snowmass G. Eigen, University of Bergen Silicon Vertex Detectors

Snowmass G. Eigen, University of Bergen L1 Trigger Rate vs Current in Machine

Snowmass G. Eigen, University of Bergen Drift Chamber Currents

Snowmass G. Eigen, University of Bergen Average Occupancy in EMC Crystals Single Crystal occupancy # Crystals with > 10 MeV  N EMC (E> 1MeV)= I HER +2.2 I LER £ N EMC (E> 10MeV)= 4.7 I HER I 2 HER +2.4 I LER I 2 LER £ with beam currents in units of [A] and luminosity in units of [10 33 cm -1 s -1 ]