SUSY-BSM meeting July 28th 2005 CERN Search for light Higgs in SUSY cascades Filip Moortgat (CERN) - Alain Romeyer (Mons – Belgium) Philip Olbrechts (CERN)
Introduction MSUGRA 5 parameters : m0, m1/2,A, tan b, sign(µ) Benchmark point LM5 Tan b = 10 AO = 0 µ > 0 M1/2 = 360 GeV M0 = 230 GeV [S. Abdullin – L. Pape Update II of the new MSUGRA test points proposal]
LSP MET B tagging h0 b Signatures ~ 20% of the SUSY events contain our decay chain … or ~ 30 000 events/year ~85% ~35% ~72%
Some “background” information … • ~ 100000 events - containing all SUSY channels - were produced at LM05. • ~ 20 % is the desired signal. • The main background contribution is expected to come from • the 80 % remaining SUSY events. • So far, no SM background events (Z+jets, W+jets, top, …) have been used. • Their contribution is expected to be small. • These will be rejected by the high jet multiplicity requirement • and the non-negligible MET cut. • The jet reconstruction is based on the “iterative cone algorithm” • (cone size: 0.5). • The combined b-tagging algorithm was applied using jets with Pt > 30 GeV. • The Jet calibration and the MET calculation are based on the “official Jet • calibration (GammaJet)” and the “EcalHcalTower” packages respectively. • DST information was read, using ORCA_8_7_3.
Analysis strategy • Event selection: Level 1 trigger, HLT, … • Reconstruction of the b invariant mass spectrum: • (combined b-tagging discriminator > 1.5) • Jet pairing is crucial, there are several posibilities: • Select the most energetic b jets. • Select the closest b jets. • Try all possible combinations of reconstructed • b tagged jets. Event with 6 b jets, two coming from the Higgs decay No simple “a priori” kinematical constraints
Resulting invariant mass spectrum Signal + Background events Signal and background minus lowest dotted red curve: highest dotted red curve Number of events Signal with two b jets, coming from the Higgs, used to reconstruct the invariant mass Signal events Background events Invariant Mass (GeV)
… and the corresponding mass resolution The width of the signal peak is mainly determined by the jet resolution of the detector as the intrinsic decay width varies from 3.3 MeV to 4.3 MeV. Number of events Mass resolution (GeV)
However, the invariant mass distribution also depends on the definition of the signal • There are mainly two ways to define the signal: • signal events are events where the process • with is identified. • B) signal events are events where the process is • identified. • As the trigger and selection criteria of the events are the same for both • definitions A and B, the global distributions (signal and background events) • are identical. However, the way in which the events are allocated as signal • or background is different …
Invariant mass spectrum (definition A) Number of events Signal events Background events Invariant Mass (GeV)
Invariant mass spectrum (definition B) Number of events Signal events Background events Invariant Mass (GeV)
Systematics: jet calibration Changing the correction factor of the “HepLorentzVector” of the jets by – 15 % and +15 % results in a shift of -18 GeV (~-16%) and + 16 GeV (~+14%) respectively. Number of events
Systematics: misalignment of tracker Applying the “short term” misalignment scenario, which corresponds to a misalignment of about 100 m, the effect on the position of the invariant mass distribution is limited (small shift to higher mass), however, the number of selected signal events drops considerably (25 – 30 %) due to the reduced b-tagging efficiency. Number of events Number of events Invariant Mass (GeV) Mass resolution (GeV)
Conclusions and future plans • The invariant mass distributions by Alain, Filip and myself – using • somewhat different approaches - are in good agreement. • A misaligned tracker has a non negligible influence on the b-tagging • efficiency, causing a reduction of selected signal events. • We need to study other systematical effects … • Calculate the significance , not easy to estimate background. • Include the ALPGEN multi-jet backgrounds. • Combine the Higgs with 1 or 2 extra jets and produce a set of • endpoint distributions to extract the sparticle masses using the • hemisphere separation method (see next talk). • Use FAMOS to study the CMS reach, i.e. scan the MSUGRA parameter • phase space so that a sensitivity region in the (m0,m1/2) plane can be • provided for several luminosities.