the photon-tuned regression. The energy needs to be smeared as well to realistically model the H → γγ invariant mass distributions.
In the H → γγ analysis, we will use the energy smearing numbers from the electron-tuned regression, because those numbers are derived from Z → e + e − sam- ples. To do this, we assume that the difference between the data and the simulation is the same for electrons and photons. This assumption is reasonable since the energy smearing is trying to account for two factors which are common to electrons and pho- tons. First, the in situ crystal-to-crystal intercalibration is taken into account in the simulation, but might be overly optimistic. This contributes to the raw supercluster energy. Electrons and photons use the same supercluster algorithm. Therefore, any miscalibration effects contribute equally to the electron or photon energy measured as the raw supercluster energy. Second, the simulation-based energy correction in data does not work as well in the data as it does in the simulation, due to the fact that not all input variables for the energy corrections are well modeled by the simu- lation. In addition, the level of disagreement between the data and the simulation for electrons and photons should be similar, since (1) nearly all variables are electromag- netic shower based and electrons and photons have similar electromagnetic behavior in the ECAL crystals, and (2) electron and photon-tuned regressions are trained with the same list of input variables and with the same number of events. Therefore we conclude that the energy smearing from electron-tuned regression, derived from Z → e + e − , is applicable to photons from H → γγ decays.
son signal. The validity of the results is checked with Z→
e + e − events.
The di-photon invariant mass resolution also depends on the opening angle between the twophotons, which in turn is related to the directions of the twophotons and the accuracy in the identification of the primary vertex where they originate. The natural spread in the longitudinal po- sition of the interaction point is about 6 cm and and the uncertainty on the position of the primary vertex worsens as the pile-up event rate increases. In CMS it was checked that if the vertex is known within 1 cm there is no im- pact on the invariant mass resolution. Techniques involv- ing multivariate methods, based on boosted decision trees, were hence developed in order to meet this target. A first multivariate analysis, MVA 1 , gives high score to likely primary vertex according to the information on the event topology extracted from all tracks present in the event and from the pointing provided from converted photons if the e + e − tracks are reconstructed. A second multivariate analysis, MVA 2 , is hence built from the score MVA 1 at- tributed to each vertex plus the γγ-system transverse mo- mentum, the number of vertices reconstructed in the event and the longitudinal distance between the highest MVA 1 - score vertex and the next two. Both multivariate analyses are trained on simulated Higgsboson events and their va- lidity is tested in Z→ µ + µ − events and in γ + jet events, the latter to verify specifically events with converted pho- tons. The primary vertex selection achieved is about 80%
The discovery of a particle consistent with the StandardModel (SM) Higgsboson in 2012 by the ATLAS  and CMS  collaborations has opened up new possibilities in searches for physics beyond the SM (BSM). Although strong astrophysical evidence [3,4] indicates the possible existence of dark matter (DM), there is no evidence yet for nongravitational interactions between DM and SM par- ticles. The interaction probability of DM particles, which are produced in SM particle collisions, with a detector is expected to be tiny. Thus, many searches for DM at the Large Hadron Collider (LHC) involve missing transverse momentum (E miss T ) produced in association with detectable particles (X þ E miss T final states). In other X þ E miss T searches in proton-proton (pp) collisions, X may represent a jet or a γ=W=Z boson, which can be emitted directly from a light quark or gluon as initial-state radiation through SM gauge interactions. However, SM Higgsboson radiation from initial-state partons is highly suppressed, so events with a final state compatible with the production of a SM Higgsboson in association with E miss T can be sensitive probes of the structure of the BSM physics responsible for producing DM. The SM Higgsboson is expected to be produced from a new interaction between DM and the SM particles [5,6] . Both the ATLAS and CMS collaborations have previously searched for such topologies using
of all the search channels considered using the CLs method [15–17] has also been presented.
2.1 H → γγ
The Higgsboson branching ratio for the decay into twophotons is approximately 2×10 − 3 between 110 and 150 GeV.The diphoton mass resolution is excellent, between 1 and 2% and the signature in this channel is two high E T isolated photons. In case of the VBF process there are two additional high p T jets that provide a further handle to discriminate the signal from the background. A signal in this channel would appear as a small and narrow peak above a large and smooth prompt di-photon background. A search is performed for a Higgsbosondecaying into twophotons. The analysis is done using a dataset recorded by the CMS experiment at the LHC from pp collisions at a centre-of- mass energy of 7 TeV, which corresponds to an integrated luminosity of 4.8 fb −1 . To improve the sensitivity of the search, selected diphoton events are subdivided into classes according to indicators of mass resolution and signal-to-background ratio. Five mutually exclusive event classes are deﬁned as shown in Figure 1: four in terms of the pseudorapidity and the shower shapes of the photons, and a ﬁfth class into which are put all events containing a pair of jets passing selection requirements which are designed to select Higgs bosons produced by the vector boson fusion process. Two photon classiﬁers are used: the minimum value of a variable R 9 of the twophotons, R min 9 , and the maximum pseudorapidity (absolute value) of the twophotons, giving four classes based on photon properties.
4 H → γγ
In spite of the small branching fraction of the Higgsboson to a pair of photons, clean signature and high di-photon mass resolution lead this channel to be one of the most sensitive ones. Di ﬀ erent production modes are exploited in this analysis and are tagged based on the presence of extra jet pair (either forward or non-forward), electron, muon or missing transverse energy. Untagged events are further categorised according to a BDT score that aims at separating categories with higher signal over background ratio from those with lower signal over BG ratio using photon identiﬁcation, isolation criteria, shower shape and mass resolution as inputs. Main backgrounds are either due to the prompt non-resonant di-photon events or events with at least one jet misidentiﬁed as an isolated photon. To suppress events from Z→e + e − events, an electron veto is applied to remove the photon candidates if they are matched to reconstructed electrons.
Abstract. One of the main targets of the CMS experiment is to search for the StandardModelHiggsboson. The 4-lepton channel (from the decay H → ZZ → 4l, l = e, μ) is one of the most promising. The analysis is based on the identiﬁcation of two opposite-sign, same-ﬂavor lepton pairs: leptons are required to be isolated and to come from the same primary vertex. The Higgs would be statistically revealed by the presence of a resonance peak in the 4-lepton invariant mass distribution. The Higgs mass is a free parameter of the StandardModel, and the 4-lepton channel search is sensitive almost in the entire mass range. With data collected in 2010 and 2011 (4.7 fb − 1 at √
3.7 H → ττ
The Higgsboson decay into τ lepton pairs is an impor- tant evidence in favour of the SM-like nature of the new boson. H → ττ events are produced either via gluon fu- sion, or WH/ZH or VBF [13, 14]. Several final states are analysed: µτ h . eτ h , eµ, τ h τ h and µµ. Isolated τ leptons hadronically decaying, τ h , are identified using a MVA al- gorithm. The ττ mass is reconstructed using the SVFit algorithm, which exploits the kinematics of the τ decay, including the invariant mass of the νν system in its lep- tonic decay, and using the transverse missing energy as a constraint. The improved mass reconstruction allows for a better discrimination between signal and the irreducible Z → ττ background. Additional contamination from other SM processes is due to QCD and W + jets.
s = 13 TeV.
The increased centre-of-mass energy leads to larger parton luminosities as shown in figure 4.7, caused the larger momentum the partons of the protons carry on average. This is the main reason for increased production cross sections at 13 TeV compared to 8 TeV. The two effects, the higher parton luminosity and the increased instantaneous luminosity lead to larger production rates for SM Higgs bosons. Moreover, the reach in sensitivity for physics beyond the StandardModel is raised as hypothetical new particles with higher masses can be produced. This sets the scope for future H → τ τ analyses. In the field of the SM Higgsboson analysis, measurements of couplings and other properties of the newly found boson can be performed with higher accuracies. As many natural extensions to the StandardModel, such as Two-Higgs-Doublet models, predict enhanced couplings of additional Higgs bosons to down-type fermions, the di-τ final state is a very important channel for probing new physics in the Higgs sector. The latter is not discussed in the scope of this thesis. However, the appendix A outlines technical developments as a preparation for covering multiple follow-up analyses.
4.2 Event Generation
The event generators model the particles involved in specific processes and the imme- diate decays in these processes. Event generation is run within the Athena framework and the results are fed into the simulation step that follows. The event generators are written and maintained by development teams external to ATLAS but Athena has a number of interface packages which allow it to drive the use of these generators. The Athena event manager generates run numbers and event numbers which are passed to the event generator which then produces an event. ATLAS uses multiple event generators and it is therefore important that all their output be mapped to a common format so that any algorithms are independent of which generator was used. The format used is the HepMC event record format . The HepMC format stores particles that are considered stable within the detector (muons are included as they live long enough to pass through the detector) to use as input to the sim- ulation step. Other particles considered as unstable in the detector are used for physics studies and diagnostics. Which particles are produced and kept depends on the generator and process being modelled. The particles are stored in a connected tree which mimics the structure of a classical decay chain. This entire connected tree is stored as the truth information in the MC datasets. Truth information gives the analyser the ability to investigate exactly what happened in the process which can be useful for checking things such as the accuracy of reconstruction algorithms. HepMC is not a standard implemented among all generators however and so Athena has to handle conversion of all different data formats into the HepMC format.
7.2 Muon reconstruction and identification
For the measurement of muons, the single most important aspect is the 3.8 T solenoidal mag- net, with inner diameter of 6 m and length of 13 m. The magnetic flux generated by the strong central field is large enough to saturate the steel in the return yoke. The standard Muon recon- struction is performed using the all-silicon inner tracker at the center of the detector and up to four stations of gas-ionization muon detectors installed outside the solenoid and sandwiched between the layers of the steel return yoke. This provides two measurements of the trajectory of each muon, resulting in a good level of redundancy and excellent momentum resolution, varying from one to few percent at 100 GeV, depending on | h | , without making stringent de- mands on spatial resolution and the alignment of muon chambers. The present detector and the favorable length to radius ratio allows efficient muon measurement up to a pseudorapid- Figure 8. Left: measured time for a particle to traverse CMS with respect to the transit time of a particle traveling at the speed of light. Muons from the decay of Z bosons are relativistic (β = 1) and HSCPs are slower. Right: the efficiency as a function of b of the standard L1 muon trigger without any pT threshold, and the RPC-HSCP Phase-2 trigger with 1.56 ns sampling time
ground; the level to which this background is controlled determines the sensitivity of the analysis.
Object reconstruction proceeds as described above.
The combination of the two highest p T jets within | ⌘ | < 2.5 is chosen as the hadronic W candidate. The events with the incorrect dijet combination comprise a broad non- peaking background in the WW mass spectrum. The lep- tonic W candidate is reconstructed from the lepton plus E miss T system. We require E miss T >25(30) GeV for each event in the muon (electron) data. In addition, the unmeasurable longitudinal component of the neutrino momentum is re- constructed by requiring the lepton-neutrino pair to have the invariant mass of a W boson. The ambiguity in the second-order equation is resolved by taking the solution that yields the smallest |p z | value.
The photon identiﬁcation eﬃciencies, averaged over η , range from 85% to above 95% for the E T range under consideration.
To further suppress the jet background, an isolation require- ment is applied. The isolation transverse energy is deﬁned as the sum of the transverse energy of positive-energy topological clus- ters, as described in Section 4, within a cone of size R = 0 . 4 around the photon candidate, excluding the region within 0 . 125 × 0 . 175 in η × φ around the photon barycentre. The distributions of the isolation transverse energy in data and simulation have been found to be in good agreement using electrons from Z → e + e − events and photons from Z → + − γ events. Remaining small dif- ferences are taken into account as a systematic uncertainty. Photon candidates are required to have an isolation transverse energy of less than 4 GeV.
2.2.5 T he E lectrom agnetic C alorim eter
The electromagnetic calorimeter is a total absorption calorimeter which detects and measures the energies and positions of electrons, positrons and photons ranging from tens of MeV to 100 GeV. It provides neutral-pion/photon discrimination and, in conjunction with the central tracking system, electron/hadron discrimination. It consists of three large overlapping assemblies of lead-glass blocks (the barrel and the two end caps). Most electromagnetic showers are initiated before the lead-glass itself because of material such as the magnet coil and the pressure vessel in front of the calorimeter. Eor this reason, presampling devices are installed in both the barrel and end-cap regions, im m ediately in front of the lead-glass to measure the positron and to sample the energies of these pre-showers^ thus improving energy resolution.
Figure 1 shows the test statistic − 2 ln Q as a function of the test mass for the LEP-wide combination. The expected curves are obtained by replacing the observed data configuration by a large number of simulated event configurations for the two hypotheses. For the background hypothesis the 68% and 95% probability bands are also shown. There is a broad minimum in the observed − 2 ln Q starting at about 115 GeV/c 2 . The negative values in this mass range indicate that the hypothesis including a StandardModelHiggsboson of such a mass is more favoured than the background hypothesis, albeit at low significance. Note also that the median expectation for the signal plus background hypothesis crosses the observed curve in this mass range. The fact that the observed curve slightly deviates from the background expectation over the whole mass range of the figure can also be explained by local upward fluctuations of the background and by long-range effects due to the experimental resolution.
pected StandardModelHiggsboson production cross sec- tion in each bin. The dependence of the signal and back- ground predictions on the systematic uncertainties is de- scribed by nuisance parameters, θ, which are parametrised by Gaussian or log-normal priors. The expected number of signal and background events in each bin are functions of θ. The parametrisation is chosen such that the rates in each category are log-normally distributed for a normally dis- tributed θ. The test statistic q µ is then constructed accord- ing to the profile likelihood: q µ = 2ln( L (µ, θ ˆ µ )/ L ( ˆ µ, θ)), ˆ where ˆ µ and ˆ θ are the parameters that maximise the like- lihood (with the constraint 0 ≤ µ ˆ ≤ µ), and ˆ θ µ are the nuisance parameter values that maximise the likelihood for a given µ. This test statistic is used to measure the compatibility of the background only model with the ob- served data and for exclusion intervals derived with the CL s method [13, 14]. The normalisation of the top, Z + b and W + b backgrounds are allowed to float freely in the fit. The other backgrounds are constrained within their er- rors as described in section 4. The resulting scale factors from the fit are shown in table 4 for both √
A search for H → b b ¯ decays is performed by looking for an excess of events above the background expectation in the invari- ant mass distribution of the b-jet pair (m b b ¯ ). The value of the reconstructed m b b ¯ is scaled by a factor of 1.05, obtained from MC-based studies, to account on average for e.g. losses due to soft muons and neutrinos from b and c hadron decays. To in- crease the sensitivity of the search, this distribution is examined in bins of p V T . As the expected signal is characterized by a rela- tively hard p T V spectrum, the signal to background ratio increases with p T V . The Z H → + − b b ¯ and W H → ν b b ¯ channels are exam- ined in four bins of the transverse momentum of the reconstructed W or Z boson, given by: p V T < 50 GeV, 50 p T V < 100 GeV, 100 p V T < 200 GeV and p T V 200 GeV. In the Z H → ν ν ¯ b b ¯ search three bins are deﬁned: 120 < p V T < 160 GeV, 160 p T V < 200 GeV and p T V 200 GeV. The expected signal to background ratios for a Higgsboson signal with m H = 120 GeV vary from about 1% in the lowest p V T bins to about 10–15% in the highest p T V bins. For this Higgsboson mass, 5.0% and 2.4% of the Z H → + − b b ¯ and W H → ν b b ¯ events are expected to pass the respective analysis selections, with negligible contributions from other ﬁnal states. On the other hand, the Z H → ν ν ¯ b b ¯ analysis has a non-negligible con- tribution from W H → ν b b: 2.1% of the ¯ Z H → ν ν ¯ b b ¯ signal and 0.2% of the W H → ν b b ¯ signal are expected to pass the analysis selection.
The search for evidence of beyond StandardModelHiggs bosons is an integral part of the Higgsboson studies at the LHC. This article reviews recent beyond StandardModelHiggsboson searches using Run I LHC proton-proton collision data recorded by the ATLAS detector. In particular, searches for Higgsboson cascades, double Higgsboson production, scalar particles decaying to γγ pairs,
The StandardModel of particle physics (SM) is a theory which describes the most funda- mental constituents of matter as well as electromagnetic, weak and strong interactions. SM predictions have been extensively tested experimentally with great success. Examples include the existence of the W and Z bosons, gluons, the top and charm quarks and most recently, the Higgsboson. Additionally, precision tests of the SM have proven to be successful with the predictions of the masses of the weak force carriers, the W and Z . However, there are several deficiencies with the SM including a lack of a dark matter candidate and a lack of a quantum field theory which accommodates general relativity. Deficiencies of the SM are explained in detail in Section 2.4.
81 University of Virginia, Charlottesville, Virginia 22901, USA
82 University of Washington, Seattle, Washington 98195, USA (Received 30 March 2009; published 25 June 2009)
We present a search for the standardmodelHiggsboson using hadronically decaying tau leptons, in 1 fb 1 of data collected with the D0 detector at the Fermilab Tevatron p p collider. We select two final states: plus missing transverse energy and b jets, and þ plus jets. These final states are sensitive to a combination of associated W=Z boson plus Higgsboson, vector boson fusion, and gluon-gluon fusion production processes. The observed ratio of the combined limit on the Higgs production cross section at the 95% C.L. to the standardmodel expectation is 29 for a Higgsboson mass of 115 GeV.
The central feature of the CMS apparatus is a supercon- ducting solenoid of 6 m internal diameter, providing a mag-
netic field of 3.8 T. Within the solenoid volume are a sil- icon pixel and strip tracker, a lead tungstate crystal elec- tromagnetic calorimeter (ECAL), and a brass and scintilla- tor hadron calorimeter (HCAL), each composed of a bar- rel and two endcap sections. Forward calorimeters extend the pseudorapidity coverage provided by the barrel and end- cap detectors. Muons are detected in gas-ionization cham- bers embedded in the steel flux-return yoke outside the solenoid.