Progress on Black Hole CSC Note M Kaneda

Progress on Black Hole CSC Note M. Kaneda On behalf of Te. V-scale Gravity group 24/Jan/2007 Outline Status Discovery analysis Systematic uncertainties Democratic decay Summary M. Kaneda 1

Status Particle ID definitions were updated as recommended by physics coordination / performance group. Most of results and figures are renewed. Systematic uncertainties about detector performance are studied. Some new contents, such as democratic decay, are added. From referees: They commented about length of our note and we agreed to cut the detector section of our note. This leaves a little over 30 pages. 24/Jan/2007 M. Kaneda 2

Data set used in our note Signal sample Dataset No. n MBH(Te. V) s(pb) 5640 2 5 -14 40. 69 5641 2 8 -14 0. 34 6640 4 5 -14 24. 32 6641 7 5 -14 22. 27 Signal samples are generated with CHARYBDIS(hep-th/0307305) All samples assume 1 Te. V Planck Mass Backgrounds sample s(pb) Dataset No. Process 5014 -5017 QCD di-jets 1. 28 x 104 5641 ttbar 8. 33 x 102 6104 -06, 6110 -2 W+jets 5. 60 x 102 8133 -5, 8145 -7 Z+jets 51. 8 6119 -22, 6125 -26 g(g)+jets 5. 07 x 103 24/Jan/2007 M. Kaneda BH events 3

Particle selection Muon Staco. Muon Combined Muon |h|<2. 5, PT>15 Ge. V etcone 20<min(100, 0. 2 Pt+20)Ge. V match chi 2 <100 Photon Egamma |h|<2. 5, PT>15 Ge. V Is. EM&0 xff=0 Jet Electron Egamma (not softe) |h|<2. 5, PT>15 Ge. V Except for 1. 00<|h|<1. 15, 1. 37<|h|<1. 52 Cone 4 Tower |h|<5, PT>20 Ge. V Is. EM&0 x 3 ff=0 For egamma eta cut, eta. BE 2 (eta of cluster, geometrical eta) is used. Overlap removal: delta. R<0. 1 with the following order m->e->g->jet 24/Jan/2007 M. Kaneda 4

Event selection Studied two different set of cuts for BH discovery study Blackhole events are generally expected as high PT multi particles events. But some parameters and models shows fewer particles decay. 1. Sum Pt Cut Σ|PT|>2. 5 Te. V At least one lepton PT>50 Ge. V 2. Multi particles Cut Number of Particles (PT>200 Ge. V) > 3 including at least one lepton. signal(5640) Sum of BG 24/Jan/2007 MBH distributions after each selection (both signal data set = 5640) MBH is calculated as: PBH = Σi=all particles Pi + (miss. ET, miss. ETx, miss. ETy, 0) MBH=√(PBH 2) M. Kaneda 5

fb-1 Discovery Potential 1. Sum Pt Cut Σ|PT|>2. 5 Te. V At least one lepton PT>50 Ge. V 2. Multi particles Cut Number of Particles (PT>200 Ge. V) > 3 including at least one lepton. MBH>8 Te. V, n=2 n=7 n=4 n=2 Above figures show a required luminosity to discover blackhole. These show that we can discover blackhole at very early stage of the experiment. 24/Jan/2007 M. Kaneda 6

Large Missing Et case Some models expect large missing ET when BH is generated. Charybdis(BH generator we are now using) can not generate such events. Right figure shows MBH distribution of BG after cut: Σ|PT|>2. 5 Te. V Miss. ET>500 Ge. V The amount of remaining backgrounds with this cut is very small with 1 fb-1. 24/Jan/2007 MBH distribution of BG with Large Miss. Et Cut M. Kaneda 7

Systematic uncertainties Uncertainties on ID efficiencies, fake rates and resolutions is listed in: https: //twiki. cern. ch/twiki/pub/Atlas/CSCNotes. List/Systematics. pdf 2 kinds of systematic uncertainties studied. (They are sensitive for BH analysis) Efficiencies of lepton Assume 1% wrong efficiencies. 1% of electron-> Jet, 1% of muon->missing Et Energy scale of Jet Assume ± 5% uncertainties Scale jet energy +5% or -5% Cross section errors of BG: https: //twiki. cern. ch/twiki/bin/view/Atlas/CSCNotes. List#Cross_sections_numbe rs_are_still 12% for ttbar 3% for alpgen samples 30% for di-jets (this was discussed in Nov TP week) 24/Jan/2007 M. Kaneda 8

Effects of systematic uncertainties Number of events with 1 fb-1 Multi particles Cut: Number of Particles (PT>200 Ge. V) > 3 including at least one lepton. 24/Jan/2007 M. Kaneda 9

Summary of systematic uncertainties Number of events with 1 fb-1 Multi particle Cut: Number of Particles (PT>200 Ge. V) > 3 including at least one lepton. Statistical errors of Monte Carlo samples is lager than other systematic errors. To reduce this errors, we should think about: how many events we should generate. what event filter we should use. This is future plan beyond CSC note 24/Jan/2007 M. Kaneda 10

Systematic errors from BH generator For signals, systematic errors depends on theoretical parameters associated with CHARYBDIS were studied. Kinematic cut Temperature variation Remnant decay Other theoretical uncertainties are also listed in the Note such as: Balding phase effect Graviton effect But these uncertainties could not be simulated with CHARYBDIS ~3% for n=2 ~80% for n=7 24/Jan/2007 M. Kaneda 11

Democratic Decay 1 A ratio of each SM particles in BH decay is one of the useful information. The ratio of e, m, Z and top is studied. In this study, background were neglected. Particle selection for this study 24/Jan/2007 M. Kaneda 12

Democratic Decay 2 Reconstructed Z(->ll) mass Reconstructed top(->bqq) mass Generator 0. 969 0. 891 0. 258 0. 290 The results are consistent with the expectations for the generated data set 24/Jan/2007 M. Kaneda 13

Statistic issues Here is one of the procedure for statistics, but we did not discuss the details in out group meeting, so here is only an idea. The cross section of BH is very large in any parameters, we can discover them up to the kinematic limit. We have to study that “what is the smallest cross section we can observe above the backgrounds, assuming that the events look as expected but there are less of them. ” A estimation of backgrounds is important. Because selections of BH are very tight, there are not enough samples of backgrounds. Some of them are no events pass the selections. We should use poisson pdf for these processes and use gaussian pdf for other case, and combine all pdfs to produce the overall pdf of background. 24/Jan/2007 M. Kaneda 14

Summary First comments of referees received and implemented. Some minor changes in particle ID definitions were needed and all results were recreated. Figures of discovery potential are added. They show that we can discover blackhole at a very early stage of the experiment. Systematic uncertainties on lepton efficiencies and jet energy scale were studied. Their effect is about a few %. Statistical errors of MC samples are dominating error currently. This is task for beyond CSC note. Democratic decay study added. Release of the final draft on 31. 01. 2008 24/Jan/2007 M. Kaneda 15