1. ATLAS UK physics meeting, 10/01/08 1 Triggers for B physics Julie Kirk RAL Overview of B trigger strategy Algorithms – current status and plans Menus Efficiencies from data Summary

2. ATLAS UK physics meeting, 10/01/08 2 B physics at LHC LHC: proton-proton collisions at √s = 14 TeV bunch crossing rate 40kHz High bb production cross section: ~500 µb (~ 1 in 100 p-p collisions → bb pair). Must select those of interest. B-physics programme includes: –CP violation (e.g. B→J/ ψ(X), B→  ) –B s oscillations (e.g. B s →D s π, B s →D s a 1 ) –Rare decays (e.g. B→  (X), B→K* γ ) –J/ ψ → μμ useful for detector alignment B-trigger is based on single and di-muons: –BR ~ 10 % –but clean signature at early level in trigger –gives flavour tag (needed in many analyses)

3. ATLAS UK physics meeting, 10/01/08 3 High luminosity (>2x10 33 cm -2 s -1 ) use di-muon trigger (p T > 6 GeV) –B → J/ ψ(  ) X –Rare decays with di-muon, e.g. B→ , B→K 0*  Low luminosity (<2x10 33 cm -2 s -1 ) add a single muon trigger with additional JET/EM ROI information from LVL1. At LVL2 have 2 possible approaches: –Full reconstruction (FullScan) inside inner detector (time costly) –Use LVL1 Regions of Interest (RoI) to seed LVL2 reconstruction: Jet RoI for hadronic final states (e.g. B s →D s (  π ) π ) EM RoI for e/  final states (e.g. J/ ψ →ee, K* γ,  γ ) Muon RoI to recover di-muon final-states in which second muon was missed at LVL1. Overview of B-trigger strategy Allows us to have a broad programme of B-physics in initial low luminosity period. Continue with rare-decay searches at high luminosity

4. ATLAS UK physics meeting, 10/01/08 4 Muon B trigger: Topological di-muon (Jpsimumu, Bmumu, Upsimumu, BmumuX) Seeded by two L2_muX X = 4GeV or 6GeV Signatures: 2MUX_Bmumu, 2MUX_Jpsimumu, 2MUX_Upsimumu Use same algo configured with different mass cuts: Bmumu: 500-8000 MeV Jpsimumu: 2600-3400 MeV Upsimumu: 8500-10500 MeV 2MUX_BmumuX –Being implemented L1_MUX  Fast  Comb L2_muX IdScan_muon EF tracking EF_muX TrigMoore L2_muX L2BmumuFex L2_2MUX_Bmumu L2BmumuHypo L2_muX EF_muX EFBmumuFex 2MUX_Bmumu EFBmumuHypo EF_muX 4 In early running also have 2muX with no further processing – for Drell Yan and cross-checks HLT muon confirmation HLT algos combine pairs of tracks make mass and vertex cuts

5. ATLAS UK physics meeting, 10/01/08 5 Muon B trigger: TrigDiMuon (Jpsimumu, Bmumu, Upsimumu, BmumuX) Seeded by single L1_MUX/L2_muX, search for second  in large RoI around the first  Signatures: MUX_Bmumu, muX_Jpsimumu, MUX_Upsimumu_FS, etc. These again use same code with different mass cuts MUX_BmumuX - being implemented L1_MUX IdScan / SiTrack L2DiMuHypo L2_MUX_Jpsimumu TrigDiMuon L2_muX EF tracking EFDiMuHypo MUX_Jpsimumu TrigMuGirl L2_muX_Jpsimumu muX_Jpsimumu 5

6. ATLAS UK physics meeting, 10/01/08 6 Di-muon B triggers Trigger J/  →  →  2MU6_Jpsimumu15% 2MU4_Jpsimumu38% MU6_Jpsimumu73% MU6_Jpsimumu_FS73% 2MU4_Upsimumu36% MU4_Upsimumu0.6% MU4_Upsimumu_FS67% Examples of efficiencies for different triggers (few 100s events) Requiring only 1 LVL1 muon and using TrigDiMuon to find the second muon with lower pT thrshold greatly increases efficiency. For Upsimumu muons are well separated and we need FullScan

7. ATLAS UK physics meeting, 10/01/08 7 Hadronic B decays B s  D s ( φ ( K + K - ) π ) π Seeded by single L1_MUX_J5, search for hadronic decay in RoI around J5 or FullScan Signatures: MUX_DsPhiPi, MUX_DsPhiPi_FS, muX_DsPhiPi, etc. Implemented at L2 and EF ! Needs work to reduce rate at higher luminosities, particularly at EF – use vertexing, reconstruct the B? L1_MUX_J5 Confirm muon at L2 L2DsPhipiHypo L2_MU(mu)X_DsPhiPi(_FS) L2DsPhiPiFex Confirm muon at EF EFDsPhiPiFex MU(mu)X_DsPhiPi(_FS) EF tracking (B physics / FullScan) IdScan / SiTrack_Bphysics / FullScan EFDsPhipiHypo 7 Combine tracks make mass cuts to find  and D s

8. ATLAS UK physics meeting, 10/01/08 8 Hadronic B decays Efficiency RoI vs FullScan Timing tests during November technical run RoI vs FullScan track reconstruction times AthenaMT (solid) vs L2PU (shaded) to study effects of real data access time (+~50%). RoI FullScan

9. ATLAS UK physics meeting, 10/01/08 9 EM B decay J/  →ee, B → K  B →  Seeded by single L1_MUX_EM3: Search for EM decay in RoI around EM3 or FullScan Signatures: MUX_Jpsie5e3 muX_Jpsie5e3_FS muX_Bgamma5x, etc. Jpsiee currently implemented at L2 For Bgamma5x: there is a chain setup with dummy hypo to allow us to perform feasibility studies. L1_MUX_EM3 Confirm muon at L2 L2JpsieeHypo L2_MU(mu)X_Jpsie5e3 L2JpsieeFex Confirm muon at EF EFJpsieeFex MU(mu)X_Jpsie5e3 EF tracking (B physics) IdScan / SiTrack_Bphysics EFJpsieeHypo 9

10. ATLAS UK physics meeting, 10/01/08 10 B → K  B →  L1: L2: “standard” low ET photon cuts from Egamma group then reconstruct  K * gives: 70% efficiency for background ~6% of L2_muon rate (~6 -25Hz) Add “extra “ calo cuts (cluster width, etc.) gives 60% efficiency background 10 Hz. Need to optimize the cuts and implement at EF

11. ATLAS UK physics meeting, 10/01/08 11 Menus for different luminosity scenarios Choose muon threshold use of RoI/Fullscan track reconstruction to optimize performance at different luminosities 10 31 : Use L1_MU4 plus tracks reconstructed by ID Fullscan. At startup have 2 muon trigger with no further selection for Drell Yan. 10 32 : Use L2_mu4 before HLT slections and use track reconstruction in Muon, Jet and EM RoI, rather than FullScan  10 33 : Raise muon threshold to mu6

12. ATLAS UK physics meeting, 10/01/08 12 Extracting efficiencies from data Don’t want to rely on MC for our trigger efficiencies Look at using data: J/  → μμ used with “tag and probe” method to get single and then di- muon efficiencies (see CSC note). Work just starting need to look at other channels. BARREL ENDCAP J/  (2.88GeV<M  <3.3GeV) seed muon (MU06) probe muon 1 2 3 Endcap Barrel –Fitting result Trigger eff. curve for Pt

13. ATLAS UK physics meeting, 10/01/08 13 Summary Much recent work to implement new trigger chains and menus. Still much to do –EF implementation in most chains –Tuning cuts (efficiency versus rate) –Efficiency studies – how to get efficiencies from data? –Monitoring – setting up histograms and automatic checking Anybody interested in getting involved please let us know.

14. ATLAS UK physics meeting, 10/01/08 14 Backup slides

15. ATLAS UK physics meeting, 10/01/08 15 Menus: Status in 13.0.30.3 Signatures in 13.0.30.3 (default menu): ["MU6_DsPhiPi", "MU4_DsPhiPi", "mu6_DsPhiPi", "mu4_DsPhiPi", "MU6_DsPhiPi_FS", "MU4_DsPhiPi_FS", "mu6_DsPhiPi_FS", "mu4_DsPhiPi_FS", "MU4_Jpsie5e3_FS", "MU4_Jpsimumu", "MU6_Jpsimumu", "MU4_Upsimumu", "MU6_Upsimumu", "MU4_Bmumu", "MU6_Bmumu", "mu4_Jpsimumu", "mu6_Jpsimumu","mu4_Upsimumu","mu6_Upsimumu", "mu4_Bmumu","mu6_Bmumu", "2MU4_Bmu4mu4","2MU6_Bmu6mu6","2MU4_Bmu4mu4x","2MU6_Bmu6mu6x", "2MU4_Jpsimumu","2MU6_Jpsimumu", "2MU4_Upsimumu","2MU6_Upsimumu"].

16. ATLAS UK physics meeting, 10/01/08 16 10 31 (Lumi1E31.py) LVL1 only: L1_2MU4 HLT passthru’ : 20Hz L1_MU4 prescale ~200 HLT pass thru’ : 5 Hz LVL1 + Muon at HLT 2mu4 : ~10 Hz mu4 prescale ~100 : 5 Hz LVL1 + ID & MU at HLT: MU4_DsPhiPi_FS, MU4_Jpsimumu, MU4_Upsimumu, MU4_Bmumu, 2MU4_Bmu4mu4, 2MU4_Jpsimumu, 2MU4_Upsimumu Note: Full-scan run at rate of either: L1_MU4 : 1kHz : 1kHz * 80ms = 80 cores = 2% LVL2 farm L2_mu4 : ~400Hz => <1% LVL2 farm Loose selections ~10Hz Aim: HLT triggers based on FullScan for all events with LVL1 MU4 Still need to implement TrigDiMuon triggers to use FullScan

17. ATLAS UK physics meeting, 10/01/08 17 10 32 (Lumi1E32.py) LVL1 only: L1_2MU4 HLT pre-scale ~100 : 2Hz L1_MU4 prescale ~5000 HLT pass thru’ : 2 Hz LVL1 + Muon at HLT 2mu4 pre-scale ~20 : ~5 Hz mu4 prescale ~2000 : 2 Hz LVL1 + ID & MU at HLT: mu4_DsPhiPi, mu4_Jpsimumu, mu4_Upsimumu, mu4_Bmumu, 2MU4_Bmu4mu4, 2MU4_Jpsimumu, 2MU4_Upsimumu Note: Note: Could run Full-scan run at: L2_mu6 : 1kHz * 80ms = 80 cores = 2% LVL2 farm Tighter selection cuts needed to achieve ~10Hz Increased pre- scales to achieve ~2Hz Require L2 mu4 and use HLT reconstuction in RoI

18. ATLAS UK physics meeting, 10/01/08 18 10 33 (not implemented yet) LVL1 only: L1_2MU6 HLT pre-scale ~100: 2Hz L1_MU6 prescale ~10000 HLT pass thru’ : 2 Hz LVL1 + Muon at HLT 2mu6 pre-scale ~50 : ~2 Hz mu4 prescale ~2000 : 2 Hz LVL1 + ID & MU at HLT: mu6_DsPhiPi, mu6_Jpsimumu, mu6_Upsimumu, mu6_Bmumu 2MU6_Bmu6mu6, 2MU6_Jpsimumu, 2MU6_Upsimumu Additional cuts needed to achieve ~10Hz (probably need decay length cuts) Raise muon threshold to mu6