Silvia Ochesanu Thomas Maes, Hans Van Havermaet University of Antwerpen October 26, 2007

1. CASTOR TRIGGER STUDY Silvia Ochesanu Thomas Maes, Hans Van Havermaet University of Antwerpen October 26, 2007

2. Outline • Background rate and signal efficiency • Drell –Yan process • Forward jets • Rapidity Gap with CASTOR/HF. • Conclusions

3. Generator study Why a generator study? • CASTOR simulation in CMSSW was completed October 9th, 2007 and was not available for this study Is a generator level study good enough? – the jet rate obtained by a generator level version of the L1 jet finding algorithm is very similar to that presented in the CMS L1 Trigger TDR →we can rely on the additional rate reduction provided by rapidity gap conditions – the CASTOR energy resolution is very good: • resolution of 1 readout unit: σE ≈ 10 → 20 GeV for E = 0 → 5000 GeV →the effect of smearing across the trigger threshold will be limited – problem: electromagnetic/hadronic energy separation →a simulation study is still needed

4. Drell Yan Process l+ p p l- if x- ≪ x+ → access to low-xBjorken proton structure; → at LHC (for Q ≳ 1 GeV and η = 8): xBjorken ≳ 10-8 → xBjorken decreases by factor 10 for each 2 units in rapidity

5. CASTOR Detector CASTOR cherenkov calorimeter ● Electromagnetic and hadronic sections ● 16-fold segmentation in φ ● 2(em)+12(had)-fold segmentation in z (readout units within one φ-sector will be grouped together for the trigger) ● No segmentation in η

6. Monte Carlo Generator • PYTHIA(6.402) used in standard mode (CTEQ 5M1 pdf) – background: minimum bias QCD (MSEL = 2)  = 101.4 mb – signal: ● select Drell Yan process (electron –positron pairs) channel, switching off gamma decays to anything except e+e-;  = 88.19 nb (2.6 nb with e+e- in CASTOR) ~ 26 million events for 10 fb- ● QCD jets (MSEL = 1)

7. Level 1 Trigger variables Definition of the variables: • Electromagnetic energy :e+, e- and  • Hadronic energy :energy sum of other particles • Hits (yes/no): charged particles present All variables calculated in 16 azimuthal segments Additional variables can be used (HLT/offline) to reduces the background: • charged multiplicity • theta , phi of charged particles • invariant mass of lepton pair Trigger : Ask for energy deposit in the electromagnetic part of CASTOR above threshold in combination with a veto for energy deposits in the hadronic part of the calorimeter.

8. Drell Yan :background rate (L=1032cm-2s-1) L1 condition: • 2 φ-sector with Eem>E0 ● Eem > 300 GeV ~ 30kHz ● Eem > 600 GeV ~ 0.6 kHz Additional conditions: • + hadronic energy <5 GeV; ~ 3kHz ~ 0.07kHz • +At least one charged particle (hits) in T2 in front of electromagnetic segment <0.01kHz

9. Drell Yan :background rate (L=1033cm-2s-1) L1 condition: • 2 φ-sector with Eem>E0 ● Eem > 300 GeV ~ 3000kHz ● Eem > 600 GeV ~ 80 kHz Additional conditions: • + hadronic energy <5 GeV; ~60kHz ~1kHz • At least one charged particle (hits) in T2 in front of electromagnetic segment <0.01kHz

10. Drell Yan :signal efficiency L1 condition:2 φ-sector with Eem>E0 ● Eem > 300 GeV ● Ehad < 5 GeV -> 20% efficienty in xBj< 5 x 10-7 for DY (no pile -up) -> lower efficiency in xBj for pile –up events

11. Background study Lower threshold to study the nature of background processes (-> high rates) • Elastic and diffractive processes hardly contribute • Background comes from QCD 2->2 processes, mostly gg ->gg • Background consist of forward energetic π0 in coincidence with a charged hadron

12. Forward jets Forward jets Forward jets are mainly produced through: qg → qg gg → gg ● Total cross section for events with highest ETjet in CASTOR (from PYTHIA): 465 μb ● Trigger: ask for large hadronic energy deposit

13. Forward jets: background rate ● L1 condition: − 1 calorimeter with Etot>E0 − 2-3 adjacent φ-sectors with Etot>E0 ● O(1kHz) rate possible for Etot > 3000 GeV (corresponding to ET≳15GeV) ● further strategies to be studied (e.g. For HLT) − require highest ETjet to be in CASTOR − cut on maximum ETof central jets

14. Forward jets: signal range xBjdistribution for events with highest ETjet (ET>5 GeV) in CASTOR: ● Weak correlation between highest ETjet rapidity and xBj ● Efficiency study suffers from low statistics

15. Rapidity Gap L = 1032 cm-2 s-1 ● Diffractive production of jets: pp → p jet jet p → possible to trigger on ET>40 GeV dijets with proton tag in L1 trigger [CMS Note-2007/002] ● No proton tags available on L1 in 2008 → need for single/double gap triggers with CASTOR and/or HF → check L1 jet rates on generator level 2.4 kHz for dijets with |η| < 2.5 and ET > 40 GeV ( 2.6 kHz CMS Note-2007/002) → multiply rate by 10! 1-jet 2-jet

16. Single Rapidity Gap (dijet+gap) with CASTOR/HF Level 1 condition: • 2 jets with |η| < 2.5 and ET> 40 GeV) ; • Etot (single sided HF/CASTOR) < E0);

17. Double Rapidity Gap with CASTOR/HFgap+dijet+gap Level 1 condition: • 2 jets with |η| < 2.5 and ET> 40 GeV) ; • Etot (single sided HF/CASTOR) < E0);

18. Conclusions • generator study of L1 rate and efficiencies for Drell –Yan , forward jets at large Δη, single and double hard diffraction • Several trigger strategies are feasible using CASTOR alone • Additional information from T2 can reduce L1 trigger rates • Still a lot of problem with background • Full simulation study is need it