FP420 Signal/Background WG

1. FP420 Signal/Background WG Andy Pilkington, Marek Tasevsky, Vojtech Turanek, Michele Arneodo, Antonio Pereira, Andrew Brandt, Arnab Pal +… (WE are not exclusive!) Periodic meetings—trying for every two weeks on Thursday Attempting to standardize cuts, MC’s, etc. (for discussion should there be approval needed to give FP420 talk? Speaker’s Bureau?) Andrew Brandt (UTA)

2. News from Marek • Vojta was trying hard to use the same tunings of MC's as Arnab but he still gets much lower reduction factors from Nc cut when applied after all other cuts. Unfortunately. Vojta has to interrupt his studies for 3 weeks or so because of his final PhD exams... • At the HERA-LHC workshop, there were a few nice talks exactly about tuning of UE for LHC which I'd like to point you to: www.desy.de/~heralhc and talks by Steve Mrenna and Arthur Moraes. They give sort of recommendation what to use and we should definitely read what the guys have done. It would be beautiful if anybody of our group could do a test with the tunings above before the FP420 TDR is finished. We discussed it already with Andy and he's willing to do this first test at hadron level. I'd be also very much eager but I'm fully absorbed by finishing the MSSM paper to be available for the FP420 TDR. • HERA/LHC: http://indico.cern.ch/conferenceDisplay.py?confId=11784

3. News from Andy I still apply two cuts on N_C^{\perp} \leq 1 and N_C \leq 5 for the background and have factor of ~100 rejection on the overlap background (it is the last cut I apply in the experimental chain). I quickly checked what happens if I lower my p_t^min of charged particles from 0.5GeV to 0.2GeV and find that the background is further rejected by a factor of 8! However, I am now in contact with ID experts in the UK to see what they think about the efficiency of track reconstruction at such a low p_T. For the purpose of your talk at UTA meeting; I am currently investigating the differences between the Pythia tunes and HERWIG/ JIMMY. Note I have just found a new HERWIG/JIMMY tune (bringing the total to three) that I will apply this weekend. By the end of next week or early the week after, I hope to have the results of the Pythia tunes (about 5 tunes) which can then be lumped onto an efficiency plot and Arnab can then choose the best and worse scenario's for his studies if he wants to. As regards the double pomeron backgrounds; I have confirmed the result by Durham that this background is negligible using the new diffractive PDF's H12006 Fit B (which is the 'best' bet according to H1 because it fits the di-jet data better than fit A). However, both A and B are unconstrained for beta (pomeron mom fraction carried by parton) > 0.8. This means that Fit A gives a good estimate of the uncertainty due to fitting. I am running Fit A this weekend to get a measure of the uncertainty. This means that Arnab should not waste his time (at first anyway) generating these events.

4. SM/MSSM Higgs production at LHC Marek Taševský (Physics Inst. Prague) HERA-LHC workshop - DESY 14/03 2007 Pile-up studies for CEP H→bb MSSM scan for CEP H→bb/WW/tautau

5. S/B for Lumi = 30 fb-1 CMS/Totem Note CERN-LHC 2006-039/G-124: Very large reduction of PU effect achieved but still not yet at S/B~1. We need another handle – try track multiplicity cut.

6. Multiplicity cuts at detector level Main idea: exploit the difference between particle multiplicities of CEP and non-diffr. processes. CEP should show rapidity gaps. Construction: only good tracks (pT,track > 0.5 GeV, 0.2 GeV), |ηtrack|<2.5 (tracker acceptance) Ncharged – Nr. of tracks outside cones of R=0.7 around axes of two lead. jets Nchargedperp – dtto + perp. to jet axes (π/3<|Φjet- Φtrack|<2π/3, 4π/3<|Φjet- Φtrack|<5π/3) Observation: the cut inefficient if tracks from all vertices counted Help: take only vertices in a z-coord. window: a) 2*σZ(beam)/√2 = 2*7.5/√2 = 10.6cm – doesn’t help too much b) 2*σZ(vtx)[timing] = 6 mm - needs timing det. c) Count just tracks associated with primary vertex! Primary vertex comes from a hard-scale event. The prim.vtx finding algorithm may not be sufficiently efficient in case of high lumi – then tracks from hard jets should help to constraint the prim.vtx better.

7. Reduction factors from Ncharged cut Two options tried: Ncharged ≤ 7 .and. Nperpcharged ≤ 1 Ncharged ≤ 6 Red.factors obtained after applying all selection cuts on signal and only jet selection cuts on bg. The two options give the same. <NPU> |Signal + PU | Non-diffr.dijets + PU 420 | comb | 420 | comb ------------------------------------------ 3.5 1.2 | 1.2 | 6 | 5 7.0 1.4 | 1.3 | 7 | 6 17.6 1.8 | 1.8 | 5 | 5 35.0 1.5 | 1.8 | 5 | 5 This is for Pythia 6.2 and default CMS tune for soft underlying event (SUE). It differs significantly from A.Pilkington result (reduction ~ 100) obtained From hadron level Herwig+Jimmy tune.

8. Enhancement in MSSM? Idea of V.Khoze, M.Ryskin, S.Heinemeyer, G.Weiglein The enhancement is evaluated using Ratio = [Γ(H→gg)*BR(H→pp)]MSSM/[Γ(H→gg)*BR(H→pp)]SM H = h, H, A; p = W, b, tau The cross section is calculated as σMSSM= σSM * Ratio σSM = σ(KMR,excl.H) * BR(H→pp)SM All MSSM quantities obtained using FeynHiggs code

9. Summary on MSSM enhancement Xsections for bb and tautau enhanced most in mhmax scenario. Xsections for WW enhanced most in small αeff scenario. Enhancement increasing with tanβ. H→bb: up to 2000 for MH~180—300 GeV and tanβ ~ 50 h→bb (tautau): up to 15 for Mh~ 115 GeV and tanβ ≈ 50 h→WW: max. 4 for Mh~120—123 GeV Discovery contour plots being finalized now – still a few tests on the optimum mass windows need to be done. Interesting discovery regions observed, especially at large tanβ.

10. Andy’s HERA/LHC talk

17. FP420 still needs theorists after all V.A. Khoze (IPPP, Durham) Hbb, higher order effects Pile-Up