Vertex Reconstruction for High Luminosity pp Running at STAR STAR-Spin luminosity requirements Method of the vertex reconstruction Application to the pileup rejection Jan Balewski, IUCF for STAR Collaboration BNL, April 8, 2000 ALICE/STAR Meeting

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting STAR Spin Program Spin Structure of the proton:   G(x) in p : direct extraction for x g  [0.01, 0.3]  via p + p   + jet +  s =200 GeV GeV qg  q    pp ~ TPC readout time 40  s leads to pileup of tracks Enhanced Luminosity Level 0 : hard photon in EMC “high tower” (tower E T >5 GeV) “trigger tower” ( E T >10 GeV in (  )=(0.2, 0.2) patch ) Level 1 :  isolation cut limit for surrounding patches Level 3 : online pileup filter

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting Events Piled up in the TPC Beam 1 Beam 2 Expected interactions per bunch crossing 110 ns bunch separation 40  s TPC readout time TPC tracks EMC calorimeter  trigger identification of tracks from the trigger event ONLINE reduction of the data volume Trigger event is piled-up with ~ min bias p+p events Enhanced Luminosity p+p STAR

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting Example of 2 Piled up Events p+p event (p T > GeV The same p+p event + 1 pileup 10  s earlier (black) Displaced vertices Color coding : green p T <500 MeV/c blue p T  [0.5, 1] GeV/c red p T >1 GeV/c Tracks NOT matched to EMC towers

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting Piled up Tracks Reconstructed in TPC TPC -250 cm 0 Z-axis B-EMC p+p event (~ 30 tracks ) screened by : 735 bunch crossings (~ 3000 tr.) during 40  s of the TPC readout time EMC records only hits (~ 20 ) from the trigger event the pileup events are displaced along the beam axis (  200 cm) E-EMC 250 cm B-EMC 200 GeV, 0.27 inter/bunch

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting Vertex Finder - Method Goal : Identify vertex position (Z V ) for the trigger event Method: reconstruct all (~3000) tracks in TPC preselect (4  3) tracks from the trigger event: # TPC clusters >20 (of 45) R XY DCA < 1.3 cm (primary track) | Z DCA | < 15 cm (diamond size) track matched with EMC (~20 ) towers find maximum of the likelihood(Z V ) defined as: (Z i DCA — Z V ) 2 ln L(Z V )   —————, i  (  i ) 2 ( numbers refer to 200 GeV, 0.27 inter/bunch ) Z-resolution depends on pseudorapidity 1  (cm) 0 01  with pileup hits without -//- truncated at |Z i DCA — Z V |> 2  i

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting Vertex Finder - Performance at  s=200 GeV Nondiffractive minb p+p  0.27 inter/bunch # of tracks Likelihood Z V (cm) Recon vertex True vertex Accidental track matched to EMC Z DCA (cm) quark-gluon Compton scattering event mixed with minb, processed by GSTAR+ TRS+TCL+TPT Found 1.3 vertices/event 85 % eve with |Z V  Z MC | < 1cm “best vertex” resolution in Z Z V  Z M-C (cm)  =2.3mm Tested with GeV,  q  [1.0, 1.3]

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting Vertex Finder at  s=500 GeV Nondiffractive minb p+p  0.77 inter/bunch Often the second interaction in the trigger bunch crossing Work in progress expected more>1 vertices/event # of tracks Likelihood True vertex 2 vertex candidates threshold 2nd interaction Z DCA (cm) Z V (cm)

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting Evaluation of the Pileup Filter Pileup Filter accepts tracks (& TPC clusters): | Z i DCA — Z V | < 4  i or matched to the EMC towers Do the preserved tracks include the physics ? Reconstructed p T (GeV/c) # tracks Tested with GeV,  q  [1.0, 1.3] Data volume reduction 1:50 50% of tracks from trigger event preserved Conclusion 1: only the low p T tracks are lost After pileup filter No pileup hits added to the events

Jan Balewski, IUCF BNL, April 8, 2000 ALICE / STAR Meeting Impact on Reconstruction of Partons Kinematics q g  q   P  in EMC   jet in TPC & EMC  R(  )<0.7 jet cone qg Conclusion 2: Tracks relevant for the reconstruction of q g  q  kinematics preserved Tested with GeV,  q  [1.0, 1.3] 0.62 Error contribution to … due to not reconstructed tracks in the TPC