1. Search for the Standard Model Higgs boson with the CMS detector Tommaso Dorigo (INFN and University of Padova) for the CMS collaboration The Large Hadron Collider Introduction The Large Hadron Collider (LHC) is a proton-proton synchrotron built 100 meters underground at the border of Switzerland and France. It has a circumference of 27 km and is instrumented with 8-Tesla superconducting magnets, providing the bending power to orbit 7 TeV protons in opposite directions. Protons in the beams travel in bunches which intersect inside the detectors at a frequency of 40 MHz. At full power about 25 hard interactions per bunch crossing are expected, posing strong demands on the triggering and data acquisition capability of the experiments. LHC is due to start a 10-month data taking run in November 2009 at the reduced c.m. energy of 7 TeV, during which less than one inverse femtobarn of collisions will be acquired by the experiments. The present study still considers the Higgs search with CMS in a scenario of 1/fb collected at the c.m. energy of 14 TeV. The search for the WW (*) decay mode at CMS[3] employs events containing exactly two opposite-charge leptons (e or  ) with transverse momenta P t >10 GeV and pseudorapidity |  | 20 GeV. The following additional pre-selection cuts are then applied: a jet veto (N jet Et>15 GeV =0), large missing energy (E t miss >30 GeV), and a dilepton mass above resonances (m ll >12 GeV). Two separate searches are performed: a cut-based analysis and a Neural-Network-based analysis (NN). In both cases the selection is optimized using the azimuthal angle between the leptons, an upper cut on the dilepton mass, and requirements on lepton momenta and missing energy. The NN analysis uses additional kinematic variables to separate the signal from the main backgrounds (see Fig. 4 below). The analyses include complete techniques to determine background rates with control samples of data. The top pair background can be sized up with events containing two additional jets, while the SM production of WW pairs can be normalized using data with m ll >115 GeV. The CL s method [1,2] is used to convert the number of expected signal and background events (Fig. 5) into a significance of the observable signal, as a function of the Higgs mass. A first evidence for the Higgs boson is likely achievable in the region of best sensitivity (see Fig. 6). The H  WW search The qqH  qq  search The CMS experiment will search for a Standard Model Higgs boson in all possible final states. The H  WW (*) and H  ZZ (*) decays provide the cleanest signatures, and might allow to obtain a first evidence of the Higgs boson already with one inverse femtobarn of luminosity. For the vector-boson-fusion channel qqH  qq  larger datasets are needed in order to reach sensitivity to Standard Model cross sections. All studies presented here consider data-driven techniques to estimate the main backgrounds, and include in the calculation of significances and limits a realistic scenario of achievable systematic uncertainties on background normalizations, Jet Energy Scale, lepton selection, miscalibrations and misalignments. Conclusions [1] U. Aglietti et al., “Tevatron-for-LHC Report: Higgs”, arXiv:hep-ph/0612172, December 17 th, 2007. [2] T.Dorigo, “ Projected exclusion limits on the SM Higgs boson by combining Higgs channels at the LHC”, poster presented at this conference, September 1 st, 2009. [3] CMS PAS HIG-08-006, “Search strategy for a Standard Model Higgs boson decaying to two W bosons in the fully leptonic final state”, January 29 th, 2009. [4] CMS PAS HIG-08-003, “Search strategy for the Higgs boson in the ZZ(*) decay channel with the CMS experiment”, February 7 th, 2009. [5] CMS PAS HIG-08-008, “Search for the SM Higgs boson produced in Vector Boson Fusion and decaying into a t pair in CMS with 1/fb”, October 22 nd, 2008. References The Standard Model (SM) Higgs boson can be produced in proton-proton collisions at the LHC by several different mechanisms (see Fig.1, right). The production by gluon-gluon fusion diagrams (see Fig.2, below) is the most frequent, with cross sections of several tens of picobarns; smaller is the production by vector boson fusion (VBF), which however provides a striking signature of two forward quark jets. The two main production mechanisms have been investigated to assess the chances of an early detection of the Higgs boson with the CMS detector. Three signatures are described below: 1) gg  H  WW (*), with the decay of both W bosons to e or  pairs; 2) gg  H  ZZ (*), with a subsequent decay of both Z bosons to e + e - or  +  - pairs; 3) qq  qqH, with a decay H   accompanied by two forward hadronic jets. Fig. 3: Schematic view of the LHC complex. Four interaction regions along the beam are instrumented with particle physics experiments: CMS and ATLAS, which are optimized for high-energy pp collisions, LHCb, which targets B hadrons in forward physics processes, and a nuclear physics experiment, ALICE. Details of the CMS detector are provided in a separate poster at this conference[2]. The H  ZZ search Fig. 2: Feynman diagram showing the mechanism of direct Higgs boson production by gluon fusion at the LHC Fig. 5: Number of events after the NN selection, for different Higgs masses. Fig. 6: Expected significance of the excess of events that can be obtained in the Neural Network search, as a function of the Higgs boson mass. Fig. 4: Output of the Neural Network for signal (black points) and backgrounds (histograms) for a Higgs boson mass of 170 GeV. An optimized cut on the NN output displayed here is used to select the final candidates. In the H  ZZ (*) search[4] events are selected to contain four charged leptons (e + e - e + e -, e + e -  +  -, or  +  -  +  - ), with pair masses M ll >12 GeV. To remove the and backgrounds further, CMS uses the combined isolation of the two least-isolated leptons, and the significance of their impact parameter with respect to the primary vertex. The reconstructed mass of the dilepton pairs is requested to lay in the windows [50-100] and [20-100] GeV. After the selection, backgrounds are almost exclusively due to SM ZZ (*) production. The four-body mass provides further discrimination, as shown in Fig. 7. Mass window cuts allow to obtain a signal-to-noise ratio larger than one throughout the 120-250 GeV Higgs mass region considered in the search. The CL s method allows estimating the significance of the signal extractable signal (see Fig. 8, right) with 1/fb of 14 TeV collision data. A sensitivity at the 2  level can be obtained for favourable mass values. A study of the vector boson fusion qqH  qq  signature of Higgs decay, for Higgs masses between 115 and 145 GeV, has been performed[5] on events containing one leptonic tau-decay candidate   e or   , collected by a low-P t electron or muon trigger. A second tau-lepton candidate from H decay is required to produce a narrow jet with E t >30 GeV and one track with P t >6 GeV within its core. The two forward jets characteristic of VBF processes provide a striking signature and are used to reduce backgrounds, mainly coming from QCD multijet production and Z   decays. The mass of two E t >30 GeV forward jets has to exceed 400 GeV, and they must be separated in pseudorapidity by more than 2.5 units. After the kinematic selection backgrounds amount to 31.8 events, with 0.6-1.6 expected from the Higgs signal, depending on M h ; a typical mass distribution is shown in Fig. 9 below. Event counts are used to determine expected limits on the signal cross section in units of the predicted SM Higgs cross section, using the CL s method (Fig. 10, bottom). Fig. 7: Effective cross section as a function of the four-lepton invariant mass after the signal selection, in the mixed ee  channel. Four different signal distributions (for masses of 130, 150, 200, and 250 GeV) are compared to residual backgrounds. The latter are essentially due to SM ZZ production (in purple) and Zbb production (in green). Fig. 8: Values of the log-likelihood ratio discriminant -2 Ln Q resulting from the combination of the three four-lepton final states studied in the search. The black curve shows the expected distribution of the discriminant in the absence of signal as a function of Higgs mass, while the red curve shows the likely observable value of the likelihood ratio in the presence of the Higgs boson. Fig.9: Reconstructed mass of the two tau- lepton candidates in 1/fb of 14 TeV data after the kinematic selection described above. The expected signal for M h =135 GeV (black empty histogram) is scaled up by a factor 10 for display purposes. Fig.10: Expected limit on the pp  qqH Higgs boson cross section in units of the SM value, as a function of the Higgs mass. The green and yellow bands show 1- and 2- standard deviation bounds in the limit that can be obtained. Fig. 1: Higgs production cross section in proton-proton collisions at 14 TeV c.m. energy, in femtobarns, as a function of Higgs mass[1]. The green curve shows the cross section of the gluon-fusion mechanism, the cyan curve the cross section of Higgs production by vector boson fusion.