Standard Model Higgs Boson Searches through the 125 GeV Boson Discovery
aa r X i v : . [ h e p - e x ] J un Standard Model Higgs Boson Searches through the 125 GeV Boson Discovery
Gregorio Bernardi and Matthew Herndon LPNHE, Universit´es Paris VI & VII, CNRS/IN2P3, Paris, France University of Wisconsin, Madison, Wisconsin 53706-1390, USA
Searches for the standard model Higgs boson are reviewed from the 2 TeV run of the Tevatron with ≃
10 fb − of recorded data, and from the 7 and 8 TeV runs of the LHC, with ≃ ≃ − ,respectively, i.e., until the July-2012 discovery of a new particle by the LHC experiments. TheCMS and ATLAS Collaborations observe independently a new boson with mass ≃
125 GeV, mainlythrough its bosonic decays in γγ , ZZ , and W + W − , consistent with the standard model Higgs boson.The CDF and D0 experiments combine their results to see evidence of a similar particle producedin association with a vector boson and decaying fermionically in b ¯ b . PACS numbers: 13.85.Rm, 14.80.Bn
Contents
I. Introduction II. SM Higgs boson phenomenology andsearch strategies
III. The Tevatron and the CDF and D0detectors b jets, and missing transverse energy7 IV. The LHC, ATLAS, and CMSexperiments
V. Tevatron low-mass Higgs boson searches
V Z with Z → b ¯ b as a test of the V H search11 B.
W H → lνb ¯ b ZH → llb ¯ b ZH → ννb ¯ b and V H → E / T b ¯ b H → b ¯ b searches 14F. Searches in τ h final states 14G. Searches in H → γγ t ¯ tH production 14 VI. Tevatron high mass Higgs boson searches
W W , W Z and ZZ VII. LHC searches in bosonic Higgs bosondecays H → γγ H → ZZ → ℓ + ℓ − ℓ + ℓ − H → W + W − → ℓ + νℓ − ¯ ν H → W W, ZZ with decays ofone boson to quarks or neutrinos 20
VIII. LHC searches in fermionic Higgs bosondecays H → τ + τ − H → b ¯ b t ¯ tH production 22 IX. ATLAS, CMS, and Tevatron Results
References I. INTRODUCTION
The origin of the masses of elementary particles, oneof the remaining puzzles of the highly successful theoret-ical model of the standard model (SM), has a potentialsolution requiring the existence of only one doublet ofcomplex scalar fields. Then the finite mass of the SMelementary fermions and bosons can be explained, afterspontaneous electroweak symmetry breaking (EWSB) ofthe originally massless Lagrangian [1–7]. This minimalapproach could be confirmed if the remnant of such abreaking, the Higgs boson, is observed with the couplingsand properties predicted in the SM. The SM Higgs bo-son does not solve all problems related to the EWSB,and is maybe only one component of the fields involved.However, its discovery would be a major step in the fi-nal validation of the SM in that it would show a unifiedapproach to the generation of boson and fermion masses.While there are other approaches to explain EWSB, noneis so far as successful as the Higgs mechanism, so we con-centrate in this review on the experimental searches forsuch a boson. With the recent observation of a new bo-son at the LHC, a major step forward has been accom-plished, but a complete validation has yet to be done.The data taking has been concluded at the Tevatron andat the LHC for the center of mass energy of 7 TeV, whilethe data are still being accumulated at 8 TeV. However,given the crucial discovery made using these searches,we provide a dedicated review based on the publicationswhich immediately followed the discovery and on the re-sults available at that time, without extending the resultsreviewed beyond discovery time.To review these milestone results, we first briefly ex-plain the phenomenology of the production and decayof the SM Higgs boson, the indirect and direct con-straints on the SM Higgs from other measurements, andthe search strategies at hadron colliders, first pioneeredat the Tevatron and then extended at the LHC. We re-view the Tevatron experiments, and their low-mass andhigh-mass analyses, then the LHC experiments, and theirsearches in bosonic and fermionic Higgs boson decays. Fi-nally we review the combinations of these searches, whichled to the discovery of a new boson by ATLAS and CMSindependently, and to the evidence, from the Tevatron,for a particle consistent with this new boson. We con-clude by briefly discussing the current knowledge on thisnew boson and on short term prospects.
II. SM HIGGS BOSON PHENOMENOLOGYAND SEARCH STRATEGIESA. Phenomenology of SM Higgs Boson production
The SM Higgs boson is a CP -even scalar, and its cou-plings to fermions and to gauge bosons are proportionalto the fermion masses, and to the squares of the bo-son masses, respectively. The effective Higgs boson-gluon coupling Hgg is dominated at leading order (LO) by aone-loop graph in which the H couples dominantly to avirtual tt pair. The much weaker effective coupling tophotons Hγγ proceeds also at LO via a loop, dominatedby a virtual W + W − pair [8]. The dominant cross sectionfor Higgs boson production is the gg → H ( gg H) process,which is known at next-to-next-to-leading order (NNLO)from perturbative calculation in quantum chromodynam-ics (QCD). This calculation is performed using a largetop-mass limit approximation and a similar calculation atnext-to-leading order (NLO) in QCD has been performedfor arbitrary top mass [9–13]. The NLO QCD correc-tions approximately double the leading-order prediction,and the NNLO corrections add approximately 50% tothe NLO prediction. NLO electroweak corrections rangebetween 0% and 6% of the LO term [14–16]. Mixed QCD-electroweak corrections O ( αα s ) are also included [17].Soft-gluon contributions to the cross sections have beenresummed at next-to-leading logarithmic (NLL), next-to-next-to-leading logarithmic (NNLL) and partial next-to-next-to-next-to-leading logarithmic (NNNLL) accuracy[18–27]. Predictions for the gluon fusion cross sections atNNLO or through soft-gluon resummation up to NNLLaccuracy and two-loop electroweak effects can be foundin Ref. [17, 27–29], including differential cross section asa function of Higgs boson transverse momentum [30, 31].Uncertainties are dominated by the modeling of partondistribution functions (PDFs) and choices of fragmen-tation and renormalization scales and are ≃
25% at theTevatron and ≃
15% at the LHC. The cross sections havealso been computed exclusively for Higgs boson produc-tion in association with one jet [32, 33] and in associationwith two jets [34][35].At the Tevatron, the next most important productionprocesses are Higgs boson production in association withvector bosons (
V H ), where V is a massive W or Z vec-tor boson. The cross sections for qq → W H or ZH are
100 125 150 175 200 225 250 275 300 m H [ GeV ] s ( pp (cid:190) → H + X ) [ f b ] Tevatron √ s (cid:190) =1.96 TeV pp – → H ( NN L O + NN LL Q CD + N L O E W ) pp – → W H ( NN L O Q CD + N L O E W ) pp – → Z H ( NN L O Q CD + N L O E W ) pp – → qq H ( NN L O Q CD + N L O E W ) pp – → tt – H ( N L O Q CD ) FIG. 1: SM Higgs boson production cross sections for pp col-lisions at 1.96 TeV [9–13, 37–49, 51–54] as functions of itsmass. [GeV] H M100 200 300 400 500 1000 H + X ) [ pb ] fi ( pp s -2 -1 L HC H I GG S XS W G H ( NN L O + NN LL Q CD + N L O E W ) fi pp qq H ( NN L O Q CD + N L O E W ) fi pp W H ( NN L O Q CD + N L O E W ) fi pp Z H ( NN L O Q CD + N L O E W ) fi pp tt H ( N L O Q CD ) fi pp FIG. 2: SM Higgs boson production cross sections for pp col-lisions at 7 TeV [55], as functions of its mass. known at NNLO for the QCD corrections and at NLOfor the electroweak corrections [36–42], with a total un-certainty of ≃ qq → qqH , which dominates over the associatedproduction at the LHC, the production cross sectionsare known at NNLO in QCD and at NLO for the elec-troweak corrections, with a total theoretical uncertainty ≃
5% [43–49], which becomes larger when exclusive re-quirements are put on the jets [50]. For the associatedproduction process ttH , the cross section has been cal-culated at NLO in QCD [51–54].The cross sections for the production of SM Higgsbosons are summarized in Fig. 1 for pp collisions at theTevatron, and in Fig. 2 for pp collisions at the LHC at √ s = 7 TeV [55, 56]. Cross sections at √ s = 8 TeV havea similar behavior but are 20-30% larger at low Higgsboson mass ( m H <
135 GeV).
B. Phenomenology of SM Higgs boson decay
The branching ratios for the most relevant decaymodes of the SM Higgs boson are shown in Fig. 3 asfunctions of m H . For masses below 135 GeV, decaysto fermion pairs dominate, of which the decay H → bb has the largest branching ratio. For these lower masses,the total decay width is less than 10 MeV. For Higgsboson masses above 135 GeV, the W W decay domi-nates with an important contribution from H → ZZ above threshold. The decay width rises rapidly, reach-ing about 1 GeV at m H = 200 GeV and 100 GeV at m H = 500 GeV. Above the tt threshold, the branchingratio into top-quark pairs increases rapidly as a func-tion of the Higgs boson mass, reaching a maximum ofabout 20% at m H ≃
450 GeV.
C. Standard model fits
While the mass of the SM Higgs boson is not givenby the theory, indirect constraints for the SM Higgs bo-son mass can be derived from fits to precision measure-ments of electroweak observables. The Higgs boson con-tributes to the observed W and Z masses through loopeffects, leading to a logarithmic sensitivity of the ratioof the W and Z gauge boson masses on the Higgs bosonmass. The top quark contributes to the observed W bo-son mass through loop effects that depend quadraticallyon the top mass, which thus also plays an important rolein the global fit. A global fit to precision electroweakdata, accumulated over the last two decades mainly atLEP, SLC and the Tevatron, gives m H = 94 +29 − GeV or m H <
152 GeV at 95% C.L. [57, 58]. Measurements ofthe top-quark mass [173 . ± . W boson mass [80 . ± .
015 GeV ([60])] were used forthese constraints. These results, compared to the alloweddirect search range from March 2012 are shown in Fig. 4.
D. Direct constraints from LEP
At the LEP e + e − collider, which operated between1989 and 2000 at √ s = 90 GeV (LEP1) or 160-209GeV (LEP2), the SM Higgs boson is generally pro-duced through Higgsstrahlung in the s channel, e + e − → HZ [61, 62], where the Z boson in the final state is eithervirtual (LEP1) or on shell (LEP2). The SM Higgs bosoncan also be produced by W W and ZZ fusion in the t channel [63–65], but at LEP these processes have smallcross sections. The sensitivity of the LEP searches tothe Higgs boson strongly depends on the center of massenergy E cm . For m H < E cm − m Z , the cross section is ofthe order of 1 pb or more, while for m H > E cm − m Z , thecross section is smaller by at least an order of magnitude.Each production and decay mode was analyzed sepa- [GeV] H M100 200 300 400 500 1000 H i gg s B R + T o t a l U n c e r t -3 -2 -1 L HC H I GG S XS W G bb tt cc ttgg gg g Z WWZZ
FIG. 3: Branching ratios for the main decays of the SM Higgsboson as functions of its mass (GeV) top m ( G e V ) W m < G e V H < m < G e V H < m
155 160 165 170 175 180 185 190 19580.380.3580.480.4580.5 top (2009), m W
68% CL (by area) m top (2012), m W
68% CL (by area) m & direct Higgs exclusion) top , m W LEPEWWG (2011) 68% CL (excluding m
March 2012
FIG. 4: Constraints from the top-quark mass, the W bosonmass and other SM measurements compared to the allowedHiggs search range in March 2012. rately. Data recorded at each center of mass energy werestudied independently and the results from the four LEPexperiments were then combined. Strong upper boundson the e + e − → ZH cross section are obtained for Higgsboson masses between 1 keV and ≃
115 GeV. The com-bination of the LEP data yields a 95% C.L. lower boundof 114.4 GeV for the mass of the SM Higgs boson [66].The median limit expected in a large ensemble of identi-cal experiments when no signal is present (simply called“expected limit” in the following) is 115.3 GeV and waslimited by the collision energy achieved by the accelera-tor.
E. Search strategies at the Tevatron
At the Tevatron, the delivered luminosity and theHiggs boson production cross section are sufficient tobe sensitive at the 95% C.L. to a Higgs boson havinga mass between 90 and 185 GeV, i.e., significant overlapwith the LEP excluded region at lower masses, and withthe LHC. Sensitivity is strongest at lower mass, 90-120GeV and around the H → W W threshold ( ≃ − H → W W decaymode at high mass, which allowed in 2009 for the ex-tension of the experimental limits on the SM Higgs bo-son mass beyond the LEP exclusion limit for the firsttime after almost ten years [67], with subsequent enlargedexcluded regions [68, 69]. These results in combinationwith the precision electroweak measurements showed, be-fore the LHC produced significant results, that a mass ofthe SM Higgs boson above ≃
145 GeV was excluded at95%C.L. [70]At low-mass, the associated production channels(
W H, ZH ) involving H → b ¯ b are the most sensitive, witha significant additional contribution from direct produc-tion with H → W W decay at Higgs boson masses as lowas 120 GeV. The
W H and ZH channels with H → b ¯ b are particularly important since this decay mode will prob-ably not be observed at the 5 σ level at ATLAS or CMSbefore additional statistics has been accumulated afterthe upgrade to the full energy foreseen for 2015. Eventhough it will be measured more precisely in the LHC ex-periments than at the Tevatron, once the full 2011-2012statistics will have been analyzed, the contribution of theTevatron results will still be significant. F. Search strategies at the LHC
The LHC was designed to have a full reach to discoverthe Higgs boson, from 0.1 to 1 TeV, which is the regionwhere it was theoretically expected to be. The high en-ergy of the LHC proton-proton collider substantially in-creases the cross section for production of a Higgs bosonvia gluon fusion ( ggH ). The cross section for vector bo-son fusion and associated production are also enhanced.Because of a more unfavorable signal-to-background ra-tio in associated production (due to the initial dominant qg or gg vs q ¯ q hard interaction), the search strategygives priority to inclusive production and the H → γγ or H → ZZ decay modes at lower masses, or simplyto the H → ZZ decay mode at higher masses. Bothmodes allow possible discovery as soon as the available in-tegrated luminosity is sufficient (of the order of 10 fb − ).Also at high mass ZZ and W + W − Higgs boson decaymodes with W and Z decays to jets or Z decays to neu-trinos have strong sensitivity. The H → W W channelis best searched in the region around the
W W on-shelldecay threshold, but does not have the mass resolutionto observe a clear resonance signal. No single channeldominates the sensitivity, so the search is also performedthrough combination of all channels. Already as of 2011this strategy has allowed for the exclusion of a Higgs bo-son between masses of approximately 130 to 500 GeV.In addition, once a new particle is observed all chan-nels are crucial to understanding its nature. In particularthe fermionic decays (in pairs of τ leptons or b quarks)are a major source of information on fermion couplings,but may require a higher luminosity. In both channels itis possible to find evidence for fermionic decays and mea-sure Yukawa couplings of the Higgs boson to the fermionsgiven SM couplings using data collected before the up-grade to the full energy foreseen for 2015. The τ modeis the stronger channel since it can be seen in all Higgsboson production mechanisms with manageable signal tobackground. Other fermionic decays will not be not ob-servable before a significant luminosity upgrade not fore-seen before several years, assuming SM branching ratios,although information is gained about the Higgs bosoncoupling to the top quark through the loop productiondiagrams. G. Simulation of background and signal processes
The general strategies used to generate backgroundand signal processes at both the Tevatron and LHC aresummarized in this section.i) High cross section backgrounds
W/Z + jets: the SMbackground processes
W q ¯ q → ℓνq ¯ q or Zq ¯ q → ℓ + ℓ − q ¯ q ,where q is used to represent light partons u, d, s andgluons ( g ), and higher cross section diboson processes,are simulated using Monte Carlo (MC) matrix elementevent generators such as ALPGEN [71],
MADGRAPH [72], and
POWHEG [73–75]. Separate samples are generated for lightparton multiplicities and in each case samples are gen-erated for each of the final state decay lepton flavors ℓ = e, µ, τ . To account for the subsequent hadronizationand development of partonic showers, the matrix eventgenerators are interfaced to PYTHIA [76] using the M.L.Mangano (MLM) factorization (“matching”) scheme [71]to remove events where soft jets developed in partonicshowering overlap phase space already covered by thematrix event generator.ii) top-antitop, “single” top and diboson production: tt production has been studied since the Tevatron run 1 andis now also studied at the LHC. The electroweak produc-tion of (single) top-quark has been observed at the Teva-tron by the CDF and D0 collaboration in 2008 [77, 78]and also more recently by the LHC experiments. Simi-larly diboson production ( W W, W Z, ZZ ) has been mea-sured at the Tevatron and LHC. Extensive studies of top,single top, and diboson production have been performedby the collaborations showing that the normalization andkinematic distributions of these processes are well mod-eled by a variety of programs including
CompHep [79, 80],
MC@NL0 [81],
MADGRAPH , and
POWHEG interfaced to
PYTHIA .The diboson processes
W W and ZZ also have substan-tial contributions from gluon-gluon initial states whichare generated using the specialized generators GG2WW [82]and
GG2ZZ [83] at LHC experiments. In cases where thereare multiple additional partons in the final state, the dataare not as constraining and the techniques listed in sec-tion i) are applied.iii) backgrounds with b -flavored quarks, W b ¯ b , Zb ¯ b :These production processes are generated using similartechniques to those described above. Normalization k-factors measured in data are applied to these associatedvector boson with b -flavored quark production processes,as their total cross sections are not precisely predicted.Scale factors to account for efficiency differences in sim-ulation and data for b -jet identification are applied, as isalso the case for other processes with b -jet in the finalstate.iv) Higgs boson signal samples: these are generatedusing programs such as PYTHIA and
POWHEG interfacedto
PYTHIA . In the case of gluon fusion production thekinematics calculated at NNLO can differ from leadingorder generation and samples are re-weighed to differen-tial NNLO calculations.For each background or signal the total cross section is normalized to the best available NLO or NNLO calcu-lations.
III. THE TEVATRON AND THE CDF AND D0DETECTORS
The Tevatron was a proton-antiproton collider whichcompleted operation in September 2011. The Higgs bo-son searches took place during run II (2002-2011) inwhich it was configured to collide beams of 36 buncheswith 1.96 TeV center of mass energy and provided anintegrated dataset of 10 fb − to the CDF and D0 exper-iments (note integrated luminosities given in this reviewrefer to integrated luminosity delivered with the detec-tors in an operational condition sufficient to be used inphysics analysis). The instantaneous luminosity reached4x10 cm s − , but the effect of the overlay of multipleinteractions remained manageable.The main components of the run II CDF and D0 de-tectors are the tracking detectors, calorimeters and muondetectors. Specific details of the CDF and D0 detectorsubsystems are available in [84] and [85], respectively,while here we briefly summarize their main characteris-tics. The kinematic properties of particles and jets aredefined with respect to the origin of the detector coor-dinate system which is at the center of the detector. Toquantify polar angles the pseudorapidity variable, definedas η = − ln tan( θ/ θ is the polar anglein the corresponding spherical polar coordinate system. A. Tracking detectors
The CDF tracking system consists of an eight layersilicon micro-strip tracker and an open-cell drift cham-ber referred to as the central outer tracker (COT), bothimmersed in a 1.4 T solenoidal magnetic field. These sys-tems combined provide charged particle tracking and pre-cision vertex reconstruction in the pseudorapidity region | η | < . | η | < . | η | < . | η | < . | η | < . B. Calorimeters
The CDF calorimeter systems are used to measure theenergy of charged and neutral particles produced in p ¯ p collisions and are arranged around the outer edges of thecentral tracking volume and solenoid. These systems con-sist of modular sampling scintillator calorimeters with atower based projective geometry. The inner electromag-netic sections of each tower consist of lead sheets inter-spersed with scintillator, and the outer hadronic sectionsare composed of scintillator sandwiched between sheetsof steel. The CDF calorimeter consists of two sections: acentral barrel calorimeter and forward end plug calorime-ters covering the pseudorapidity region | η | < .
64. Thecalorimeters can identify and measure photons, jets frompartons, missing transverse energy, and in combinationwith information from other systems electron and tauleptons.The D0 liquid-argon calorimeter system is used forthe identification and energy measurement of electrons,photons, and jets, and also allows the measurement ofthe missing transverse energy ( E / T ) of the events, typi-cally from unobserved neutrinos. The central calorimeter(CC) covers detector pseudorapidities | η | ≤ . | η | = 4 .
2. They are located outside of the tracking andsolenoid systems. The calorimeters are subdivided intoelectromagnetic (EM) followed by fine hadronic and thencoarse hadronic sections. The intercryostat plastic scin-tillator detectors complete the calorimeter coverage inthe intermediate pseudorapidity region 0 . < | η | < . C. Muon detectors
The CDF muon detector is made up of four indepen-dent detector systems outside the calorimeter modulesand consists of drift chambers interspersed with steel lay-ers to absorb hadrons. The central muon detector (CMU)is mounted directly around the outer edge of the centralcalorimeter module and detects muons in the pseudora-pidity region | η | < .
6. The central muon extension iscomposed of spherical sections and extends the pseudo-rapidity coverage in the range 0 . < | η | < .
0. Thecentral muon upgrade (CMP) surrounds portions of theCMU and central muon extension (CMX) systems cover-ing gaps in angular coverage and allowing excellent iden-tification of higher momentum muons due to additionallayers of steel absorber. The barrel muon upgrade (BMU)is a barrel shaped extension of the muon system in thepseudorapidity region 1 . < | η | < .
5. The CMX, CMPand BMU systems also include matching scintillator sys-tems which provide timing information to help identifycollision produced muons.The D0 muon detector system consists of a centralmuon detector system covering the range | η | < < | η | < D. Triggering systems
The CDF trigger system consists of three levels. Levelone trigger hardware consists of dedicated electronicsthat operate at the beam crossing frequency. The levelone trigger can identify and measure the transverse mo-mentum of charged particles using COT information andbe combined with information from the calorimeters ormuon systems to provide a trigger for leptons. Thecalorimeter trigger hardware measures energy clusterswhich are used to identify jets and photons as well asan imbalance in event transverse energy interpreted as E / T . The second level trigger hardware at CDF refinesthe measurements of the level one trigger at higher preci-sion. The level two trigger can also include tracking andvertexing information from the silicon detectors. Thethird level of the trigger operates on commercial com-puters (PCs) and executes fast versions of the full offlinereconstruction software.The D0 trigger system also has three trigger levels re-ferred to as L1, L2 and L3. Each consecutive level re-ceives a lower rate of events for further examination. TheL1 hardware based elements of the triggers used in theelectron channel typically require calorimeter energy sig-natures consistent with an electron. This is expandedat L2 and L3 to include trigger algorithms requiring anelectromagnetic object together with at least one jet forwhich the L1 requirement is calorimeter energy deposi-tions consistent with high- p T jets. For muon samples,events are triggered using the logical .OR. of the full listof available triggers of the D0 experiment. The muontrigger pseudorapidity coverage is restricted to | η | < . W +jet events ( ≃ p T muons at L1. Eventsnot selected by the high- p T muon triggers are primarilycollected by jet triggers. E. Physics object identification at the Tevatron
1. Lepton identification
Isolated electrons are reconstructed in the calorimeterand are selected in the pseudorapidity regions | η | < . | η | < . . < | η | < . R = p (∆ η ) + (∆ ϕ ) > .
2. Jets, b jets, and missing transverse energy In CDF, jets are reconstructed using a calorimeterbased clustering algorithm, with a cone of size ∆
R < .
4. In D0, jets are reconstructed in the calorimetersfor | η | < . R < . γ +jet events.Jet identification efficiency and jet resolutions are ad-justed in the simulation to match those measured in data. At high instantaneous luminosity, the jets are further re-quired to contain at least two tracks with p T > . b quark (“ b tagging”) is done in two steps [87].The jets are first required to pass a taggability require-ment based on charged particle tracking and vertexinginformation, to ensure that they originate from the inter-action vertex and that they contain charged tracks. AtD0 a b tagging NN is applied to the taggable jets. ThisNN uses a combination of seven input variables, five ofwhich contain secondary vertex information; the numberand mass of vertices, the number of and χ of the ver-tex contributing tracks, and the decay length significancein the x − y plane. Two impact parameter based vari-ables are also used. At CDF the next step in b taggingis done using a NN with similar variables but includingadditional track quality information [88]. The CDF ex-periment also employs a cut based secondary vertex tag-ger [89]. As an example at D0 the typical efficiency foridentifying a p T = 50 GeV jet that contains a b hadronis (59 ± u, d, s, g ) initiated jets. This oper-ating point is typically used for events with two “loose”( L ) b -tagged jets. When tightening ( T ) the identificationrequirement, the efficiency for identifying a jet with p T of 50 GeV that contains a b hadron is (48 ± E / T )is calculated fromindividual calorimeter cell energies in the calorimeter. Itis corrected for the presence of any muons and all energycorrections to leptons or to the jets are propagated to E / T . Both experiments identify events with instrumen-tal E / T by comparing missing transverse energy calcula-tions based on either reconstructed tracks or calorimeterdeposits. The CDF experiment employs an algorithmthat combines tracking and calorimeter information toimprove E / T resolution. IV. THE LHC, ATLAS, AND CMSEXPERIMENTS
The LHC accelerator is a proton-proton collider op-erating at the highest energies currently attained by ahadron collider, which started operation in 2010. Duringthe 2011 data taking period it was configured to collidebeams with 7 TeV center of mass energy and provided anintegrated dataset of approximately 5 fb − to the LHCexperiments, while during 2012 it was configured at 8TeV and provided an integrated dataset of approximately5.8 (5.3) fb − to ATLAS(CMS). The instantaneous lu-minosity of the LHC has reached 7.7x10 cm s − mak-ing 20-30 multiple interactions per crossing a typical oc-currence and creating additional challenges for triggeringon and reconstructing physics events. A proton-protoncollider at high energy provides large cross sections forgluon-gluon or quark-quark initiated Higgs boson pro-duction processes such as gluon fusion and vector bosonfusion. For instance, the cross section for gluon fusion toa Higgs boson is increased by a factor of approximately15 compared to the Tevatron. In addition, the increasein center of mass energy from 7 to 8 TeV correspondinglyraises the Higgs boson cross section by an additional fac-tor of approximately 30%.The LHC experiments are forward-backward and cylin-drically symmetric detectors with tracking, calorimetricand muon detector elements.The ATLAS detector includes an inner trackingand vertexing system, electromagnetic and hadroniccalorimetry and an outer muon detection system [90].The inner tracking detector consists of a silicon pixeldetector, a silicon micro-strip detector, and a transi-tion radiation tracker immersed in the field of a 2 Tsolenoidal magnet which provides charged particle track-ing and vertex finding over a large pseudorapidity rangeof | η | < .
5. The inner tracker and solenoid is surroundedby a high-granularity liquid-argon sampling electromag-netic calorimeter which provides electron (photon) find-ing in the range | η | < .
47 ( | η | < . | η | < .
9. The muon spectrometersurrounds the calorimeters and consists of three large su-perconducting toroids, each with eight coils, a system ofprecision tracking chambers, and detectors for triggeringin the range | η | < . b tagging over a large pseudorapidity rangeof | η | < .
5, which is well matched to the coverage of thebarrel and electromagnetic calorimeter as well as thatof the muon chambers providing coverage to | η | < | η | < . τ lep-tons in hadronic decay modes is performed within theoverlapping pseudorapidity range of the tracker and elec-tromagnetic calorimeter. Jet finding can be performed inan expanded pseudorapidity range up to | η | < . A. Tracking detectors
The ATLAS tracker consists of a silicon pixel and striptracker and a transition radiation straw tube tracker im-mersed in a 2.0 T magnetic field provided by a solenoidalmagnet. The strip tracker consists of four barrel layersand nine end-cap disks at each end. The pixel detectorconsists of three barrel layers and two end-cap disks ateach end. The silicon inner detector tracks charged parti-cles over the range | η | < . b -jet identification. The transition ra-diation tracker consists of 4 mm drift tubes configured ina barrel region and two sets of multiple wheel endcaps.This configuration provides high efficiency r − φ trackingwith typically 36 hits per track in the range | η | < . | η | < . b -jet identification. B. Calorimeters
The ATLAS calorimeter uses sampling technologiesover the entire angular area of coverage. The electro-magnetic calorimeter is a lead-liquid argon (LAr) sam-pling calorimeter and is divided into overlapping barreland end-cap portions with coverage up to | η | < .
2. Thecalorimeter is divided into three barrels and two end-cap parts, with very fine granularity in their inner lay-ers providing excellent angular resolution for electronsand photons. The barrel calorimeter has an extra in-ner LAr layer that functions as a preshower detectorwith high longitudinal segmentation. The mass resolu-tion for diphotons with an invariant mass of 125 GeV isapproximately 1.5 GeV. The system has sufficient angu-lar pointing ability to loosely associate photons with aprimary interaction vertex to reduce combinatoric back-grounds. The hadronic sampling calorimeter is dividedinto a barrel region and endcaps that cover the range upto | η | < . | η | < . | η | < . in the barrel and 29x29 mm in the endcap, and thecalorimeter material has a Moliere radius of 21 mm lead-ing to narrow showers and good angular resolution. Theend-cap calorimeter has two layers of silicon detectorsinterleaved with lead layers configured as a preshowerdetector. The excellent energy and good angular reso-lution of the calorimeter gives a diphoton mass resolu-tion of 1.1 GeV at a mass of 120 GeV. The hadronicsampling calorimeter is divided into a barrel region andendcaps that cover the range up to | η | < . | η | < . C. Muon systems
The ATLAS muon system is based on the magneticdeflection of muon tracks within large superconductingair-core toroid magnets, instrumented with separate trig-ger and high-precision tracking chambers. The magneticfield is generated by three toroids, one in the barrel re-gion, | η | < .
4, and two in the end-cap regions, 1 . < η < .
7, with an overlapping region in between. This magnetconfiguration provides a field which is mostly orthogonalto the muon trajectories over the entire η range. Preci-sion tracking is provided by monitored drift tubes andcathode strip chambers, while resistive place chambersand thin gap chambers provide fast triggering capabilitybased on independent measurements of the particle mo-mentum for the central and forward regions respectively.The CMS muon system makes use of three technologiesinterleaved in the steel return yoke of the magnet; drifttubes (central) and cathode strip chambers (forward) forprecision tracking and triggering based on tracking infor-mation and resistive plate chambers for tracking and trig-gering based on time measurements. The return field ofthe solenoid saturates the return yoke providing a bend-ing field for independently measuring muon momenta.The chambers are divided into four stations with 14 lay-ers of drift tubes or six layers of cathode strip chambersfor robust tracking. D. Triggering
Both detectors use a multilevel triggering system. Thefirst level of the trigger allows for the measurement of themomentum or energy of physics objects including elec-trons and photons as electromagnetic energy deposits,muons with independent measurement of the momentumin the muon systems, jets and missing transverse energyusing full calorimeter information, and taus as narrowjets. The event rate is reduced to approximately 100kHz at the level one. The ATLAS experiment employsa second level trigger which repeats physics object iden-tification using the full granularity of the detector andfurther reduces the event rate to 3.5kHz. Both detectorsemploy a full event reconstruction using optimized ver-sions of offline reconstruction code running on commer-cial processors as a level three or high level event filterwith an output rate of about 200 Hz for ATLAS and 500Hz for CMS.
E. Physics object identification at the LHC
Both experiments classify observed signatures in theirdetectors as physics objects that can be associated withthe particles and decay products of particles produced inhigh-energy collisions. Physics analysis can be performeddirectly on these physics objects. The various physics sig-natures identified by the experiments are discussed next.
1. Charged lepton identification
For electron identification both experiments use clus-ters formed in the electromagnetic calorimeters and as-sociate them to tracks found in the tracker by match-ing to their extrapolated position and energy [92–94].The clustering algorithm takes into account the typicalspread of the cluster in ϕ due to bremsstrahlung pho-tons. Tracks are generally identified using inside-outalgorithms since the inner pixel detectors are the leastoccupied tracking system due to their high granularity.Electron candidates are required to pass requirements oncluster shape information, energy leakage information inthe hadronic calorimeter, and track quality information.Different operating points are defined for different levelsof selection efficiency and background rejection. Tighteroperating points with lower efficiency and better back-ground rejection include tighter criteria on identificationrequirements, requirements to reject electrons from pho-ton conversions, and requirements on track impact pa-rameter to reject electrons from interactions with matteror long=lived decays. The CMS experiment employs aBDT based multivariate electron identification algorithmfor several analyses to improve efficiency and backgroundrejection.Because of the large bending fields in the muonspectrometers of the LHC experiments, muons can be0identified in the muon systems independently of thetracker [94–96]. In addition muons can be identified inthe combined muon and tracking systems. Both exper-iments form combined tracker and muon system muonsby associating independently reconstructed muons in themuon systems to charged tracks in the tracker using po-sition and momentum information. CMS additionallyidentifies muons by extrapolating tracks into the muonsystem to perform an inside-out search for compatiblemuon system hits, which improves muon finding effi-ciency at low transverse momenta. Muons used in analy-sis are generally required to be identified in both systems,pass minimum tracker hit and muon system segment re-quirements, and satisfy requirements on track impact pa-rameter to help reject decays to muons from long livedparticles and cosmic rays.Both experiments increase the purity of identified elec-trons and muons by requiring that the charged leptoncandidates be isolated, which rejects real and misrecon-structed charge leptons contained within jets. The AT-LAS experiment applies relative calorimeter based iso-lation for electrons and relative tracker based isolationfor muons based on the total calorimeter energy or trackmomenta found in a cone around the candidate dividedby the transverse energy or momentum of the candidate.The CMS experiment applies relative particle flow basedisolation based on charge tracks; electromagnetic energyfrom electrons, photons or neutral pions; and neutralhadronic energy not associated with tracks found in acone around the candidate and divided by the transverseenergy or momentum of the candidate. These ratios arerequired to be less than a given value which can be ad-justed to achieve different levels of performance. Theparticle flow technique as used at CMS is described inmore detail at the end of this section.The identification of hadronically decaying τ leptonsis characterized by the presence of one or three chargedhadrons, identified as tracks with associated calorime-ter energy, and possible narrow strips of electromagneticenergy deposits characteristic of neutral pion decay tophotons, all contained in a narrow collimated jet [97].The ATLAS experiment combines this information to-gether in a boosted decision tree based multivariate dis-criminant [98]. The CMS experiment uses a particle flowtechnique to measure the charged hadrons in the trackingdetector and neutral pions as strip shaped electromag-netic energy deposits. It also improves the mass resolu-tion of objects reconstructed as a hadronically decaying τ by performing a fit constraining the objects from the τ decay to the τ lepton mass.
2. Photon identification
The identification of photons uses similar criteria tothose used for electrons except that events with a trackor track segment compatible with the electromagneticcluster are rejected. Photon candidates are formed from electromagnetic clusters in the EM calorimeter [99]. Theclustering algorithm and subsequent identification crite-ria allows for the possibility that the photon converts toan electron pair. Photon candidates are required to passrequirements on cluster shape information and energyleakage information in the hadronic calorimeter. TheATLAS experiment additionally applies isolation require-ments, while the CMS experiment includes the abovecriteria and isolation information in a BDT algorithmdesigned to reject non prompt sources of photons. TheATLAS experiment also uses the longitudinal segmenta-tion of its EM calorimeter to require that the photonsare compatible with pointing to the primary high trans-verse momentum interaction vertex. In cases of conver-sion early in the material of the tracker, both experi-ments reconstruct the electron-position conversion pairswhen possible. In this case the vector sum of the trackmomenta are required to point toward the primary inter-action vertex.
3. Light and heavy flavor jets
Jets are generally reconstructed using the anti- k t algo-rithm based on calorimeter clusters [100]. The excellentgranularity of the LHC detectors allows for the effectiveuse of such an iterative clustering-based jet-finding algo-rithm. Raw jet energy measurements are corrected forimperfect calorimeter response using correction factorsfrom studies of detector response in test beam data, MCsimulations, and collision data. In high pileup conditionsjets can be required to point to the hard interaction ver-tex. Heavy flavor jets, or jets originating from b quarks,are found based on the positive track impact parame-ter significance of tracks and reconstructed secondaryvertices [101, 102]. The ATLAS experiment combinesthe track impact parameter significance information toform a likelihood ratio quantifying whether a jet origi-nates from a b -flavored parton or light parton. A secondlikelihood is formed using secondary vertex informationincluding decay length significance and vertex mass infor-mation. The further use of a likelihood or similar tech-nique allows the results of the two types of algorithmsto be combined into a single continuous b -flavored jetidentification variable. The experiments define multipleoperating points with different selection efficiencies andbackground rejection. As an example using this type ofinformation both experiments can achieve 50% efficiencywith a factor of 1000 in background rejection.
4. Missing transverse energy
Both experiments use measurements of missing trans-verse energy to identify events with neutrinos [103, 104].The ATLAS experiment measures the visible energy us-ing electromagnetic clusters corrected for the averagehadronic component and corrected for the transverse mo-1mentum of identified muons. The magnitude and direc-tion of the missing transverse energy are obtained fromthe energy imbalance in the transverse plane. The CMSexperiment uses a a particle flow method to measure thevisible energy and infers the E / T measurement in the sameway. The E / T measurement can be improved in the pres-ence of high pile-up by using associated objects to cal-culate the E / T for a given vertex. The excellent trackingefficiency and z coordinate resolution of the experimentsare essential for this technique. In addition events withfalse E / T from jet mismeasurement are identified by meth-ods such as checking whether the E / T is collinear withjets or charged leptons or comparing different methodsof the E / T measurement such as those based solely on re-constructed charged tracks or calorimeter energy clusterinformation.
5. Particle flow
In CMS, jet and missing transverse energy recon-struction, and τ lepton identification are substantiallyimproved by using particle flow identification tech-niques [105, 106] that classify detector signatures ascharged or neutral hadrons, photons, or charged leptonsusing combined information from the tracker, calorime-ters, and muon detectors. Electrons, photons, and neu-tral pions are measured in the EM calorimeter. Muonsare measured in the tracker and muon detectors. Tauleptons in decay modes involving hadrons are found com-bining tracker information for charged hadrons and EMinformation for neutral pions. Charged hadrons in jetsare measured using the tracker. Neutral energy not asso-ciated with any of the above objects is measured as en-ergy clusters in the EM and hadronic calorimeters. Thevector sum of particle flow objects can also be used toidentify missing transverse energy. V. TEVATRON LOW-MASS HIGGS BOSONSEARCHES
At lower masses ( m H <
135 GeV), the dominant de-cay of the Higgs boson is H → b ¯ b , but it is hopelessto search for direct ( gg → H ) Higgs boson productionin this decay mode, due to the overwhelming multijetbackground. However, q ¯ q annihilation results in associ-ated vector boson-Higgs boson production ( V H ) in p ¯ p collisions with a better signal-to-background ratio thanavailable at a pp collider. The “primary” channels forsearching for a low-mass Higgs boson at the Tevatron, W H and ZH production, are best studied in the ℓνb ¯ b , ℓℓb ¯ b or ννb ¯ b final states. The lower branching ratios orpoor signal to background of the other decay modes ren-der their sensitivity smaller than these primary channelsbut are also searched for to provide additional sensitivityin the combination of all channels.The associated production analyses generally proceed in three steps. The first is preselection where a highstatistics sample of events containing bosons and jetsis constructed, allowing for detailed validation of back-ground modeling. The W and Z are required to decayleptonically [topologies where W or Z decay hadronicallyare also searched for, but are less sensitive ([107])] to facil-itate event triggering and selection. Electrons and muons(including those coming from taus decaying leptonically)allow for a relatively pure selection, but Z → νν de-cays are also exploited. In the next selection step, atleast one jet is required to be identified as a b -quark jet,enhancing the signal to background in separate final sub-samples defined by the number and type of b -quark jetsfound. In the final step, multivariate analysis techniquesare applied to each of the samples to further separate thepotential signal from the backgrounds.Vector boson fusion production, associated productionwith vector bosons, and direct Higgs boson productioncan also be exploited when the Higgs boson decays to a τ τ pair, by making use of the kinematics of the potentialadditional jets in the final state. These processes suf-fer from significant background and so are considered assecondary channels at the Tevatron. Another secondarychannel which is exploited at the Tevatron is inclusiveproduction of a Higgs boson with Higgs boson decay totwo photons. The sensitivity of this channel is low due tothe small branching ratio of this decay, typically smallerthan 0.2 %. The t ¯ tH production is also searched for, buthas low sensitivity.In all the analyses, the data are separated into mul-tiple orthogonal search samples of varying sensitivities.The analyses are described next and all results are sum-marized in section IX. A. V Z with Z → b ¯ b as a test of the V H search
Since the low-mass analyses use advanced multivari-ate techniques for separating signal and backgrounds andare obtained from low signal-to-background search sam-ples, a crucial test has been performed considering
W Z and ZZ diboson production with Z decays to heavy fla-vor as signal, mimicking the final states of W H and ZH including the essential feature of resonant dijet produc-tion. In such analyses the W W process is taken as abackground, with a normalization constrained to its NLOcross section. These analyses, and their combination, areperformed in the same way as their Higgs counterparts.The CDF+D0 combination displays strong evidence (4.6 σ ) for such production, with a measured cross section σ ( V Z ) = 4 . ± .
97 pb consistent with the SM predic-tion [108]. Evidence was first seen by both collaborationsseparately [109, 110] using 9.5 and 7.5 fb − respectively.2 B. W H → lνb ¯ b The search for the process q ¯ q → W H + X in which aquark-antiquark pair leads to the production of the Higgsboson in association with a W boson is based on a totalintegrated luminosity L ≃
10 fb − of collision data col-lected by both the CDF and D0 detectors at the FermilabTevatron p ¯ p collider between 2002 and 2011 [111–114].Candidate W boson events are preselected via their de-cays to an electron or a muon plus a neutrino ( W → e or µν ) while the Higgs boson is identified through its decaymode into a pair of b quarks ( H → b ¯ b ). The experimen-tal signature is a single isolated lepton, missing transverseenergy, and either two or three (to accommodate addi-tional gluon radiation in the hard collision) jets, at leastone of which is required to be consistent with having beeninitiated by a b quark.To increase signal acceptance, the lepton identificationcriteria are as loose as possible. This results in back-grounds originating from MJ events, in which one of thejets is misidentified as an isolated lepton. In the CDFanalysis, the MJ background is strongly reduced by kine-matic cuts and by using a dedicated multivariate tech-nique to reject this background [111]. The remaining MJbackground contribution is modeled from the data usingside-band techniques. In the D0 analysis, the MJ back-ground contributions passing the preselection criteria ineach sample are determined from the data using an un-binned matrix method approach [112]. The “physics”backgrounds with similar event topologies are modeledusing Monte Carlo event generators. The SM predic-tions are used to set the relative normalizations of allof the generated samples, with additional normalizationfactors applied to samples of W bosons + n partons gen-erated using the ALPGEN
Monte Carlo event generator.These factors are determined at the preselection stagewhere the SM Higgs boson contribution is negligible. Thepredicted backgrounds model the data well in the highstatistics sample before a b -tagging algorithm is applied,as shown in Fig. 5.The CDF and D0 analyses proceed by subdividing theselected sample into orthogonal subsamples based on howmany of the jets in the event, one or two, are consis-tent with having been initiated by a heavy b -quark, andat what level (“loose ( L )” or “tight ( T )”) of confidence.CDF has five tagging categories ( T T , T L , LL , T , L ) for thetwo jet sample, and two categories ( T T , T L ) for the threejet sample, while D0 uses four categories (
T T , T L , LL , T )and two categories ( LL , T ) respectively. In two b -taggedjet events, the dominant remaining backgrounds are from W b ¯ b , t ¯ t , and single top-quark production. In single b -tagged jet events the dominant backgrounds are W +light or c -quark jet production as well as MJ backgroundevents. To further discriminate the remaining back-grounds from the signal, MVA techniques are applied toeach subsample. Some of the discriminant variables usedin these analyses are given in Table I.Systematic uncertainties affect not only the normal- Dijet Mass (GeV)
500 100 150 200 250 300 350 400 E v en t s / G e V -1 D , 9.7 fb · ( DataMultijetV+lfV+hfttsingle tVV =125 GeV H M)+2 jets, Single and Double Tags nfi W( (a)
FIG. 5: Dijet mass distribution in the D0
W H analysis forthe W + 2 jet sample with 1 or 2 identified b -jets. The dataare well described by the sum of all the SM backgrounds. Thesimulated signal is also represented.rf input variable Description E / T Missing transverse energy M TW Lepton- E / T transverse mass p T ( ℓ - E / T system) p T of W candidate p T ( j ) p T ( j ) Leading (subleading) jet p T m jj Dijet invariant mass p T (dijet system) p T of dijet system∆ R ( j , j ) ∆ R between jets∆ φ ( j , j ) ∆ φ between jets H T Scalar sum of p T of all jetsTABLE I: Description of some characteristic kinematic inputquantities of the MVA technique used in the W H analyses. ization of the signal and backgrounds, but also the shapeof the MVA output distributions. The influence of eachsource of systematic uncertainty is studied separately foreach of the independent subsamples. Uncertainties inthe efficiencies of selection, on jet calibration, and onthe b -tagging criteria affect the precision at which thebackground modeling is known. The uncertainties onthe parton density functions and the effect of renormal-ization and factorization scales on signal and backgroundsimulation are also taken into account. All these uncer-tainties are allowed to affect the shape of the MVA outputdistributions. C. ZH → llb ¯ b The search by the CDF and D0 collaborations at theTevatron for the process q ¯ q → ZH + X in which a quark-antiquark pair leads to the production of the Higgs bo-son in association with a Z boson decaying to a pair ofcharged leptons is also based on a total integrated lumi-nosity L ≃
10 fb − [115–118]. Candidate Z boson eventsare preselected via their decays into e + e − or µ + µ − pairs,and the associated Higgs boson is identified through itsdecay into a pair of heavy b -quarks ( H → b ¯ b ). Candidateevents are required to have two or three jets, at least oneof which is identified as a b jet.3In this final state, which requires two leptons, the MJbackground is negligible. The physics backgrounds aremodeled using the same Monte Carlo event generatorsused in the W H analysis.To maximize the lepton acceptance and benefit fromhigher quality lepton categories, the events are classi-fied according to the lepton types. Those having bothleptons identified with high confidence are treated sep-arately from the others which contain loosely identified,forward, or track-based leptons. These samples are an-alyzed independently, allowing for an optimal sensitivityof the search. In addition, multivariate lepton selectionsare used. In CDF, to enhance the discriminating powerof the dijet invariant mass, a NN derived energy correc-tion is applied to the jets. This correction depends on themissing transverse energy and its orientation with respectto the jets. In D0, jet energy resolution improvements areobtained through a kinematic fit of the complete event,since all particles can be detected in this process.These analyses also proceed by subdividing the se-lected sample into orthogonal subsamples based on thenumber and the quality of the b -tagged jets in the event.CDF has four tagging categories ( T T , T L , LL , T ) for boththe Z + 2 and Z + 3 jet samples, while D0 has a differenttreatment using two tagging categories ( T L , T ) for the Z + 3 jet sample. CDF employs two NNs to simultane-ously separate signal events from the dominant Z +jetsand kinematically different t ¯ t backgrounds. These NNsuse various kinematic distributions, matrix element prob-abilities, and the output of a separate jet flavor sep-arating NN as inputs. In single and double b -taggedjet events, the dominant remaining background is Zb ¯ b .To suppress the remaining background MVA techniquesare applied to each subsample. Systematic uncertaintiesare overall less important than for W H since no missingtransverse energy is involved, but most of the other sys-tematic uncertainties are of similar magnitude to thoseof the
W H analyses. D. ZH → ννb ¯ b and V H → E / T b ¯ b The remaining b ¯ b analysis is built to detect the ZH → ννb ¯ b process but is also sensitive to W H events in whichthe charged lepton is not identified, hence its alternatelabel as
V H → E / T b ¯ b . These searches are based on a totalintegrated luminosity L ≃
10 fb − [119–122]. Since thisfinal state contains no leptons, triggering on these eventsand modeling the effects of the trigger requirements onthe event selection are significant challenges. Both CDFand D0 use triggers based on E / T , with or without accom-panying jets. The analyses are performed while studyingin parallel several control samples to monitor the under-standing of the background. Events are required to havesignificant E / T and two or three jets, well separated fromthe E / T direction. For the preselection, multivariate ap-proaches are also applied to the events to remove a largepart of the MJ background. For the final selection b tag- SIG NN E ve n t s / b i n [f i t t o d a t a ] : SecVTX + SecVTX (SS) -1 +b-jets 9.45 fb T E Diboson W/Z + h.f. Top MultijetData (x5) Higgs 125 GeV/c [CDF II Preliminary]Signal region
FIG. 6: Distribution of the NN discriminant in the CDF E / T bb analysis for the sample with two or three jets where2 of the jets are tagged as b jets by the secondary vertexalgorithm. The data are well described by the sum of all theSM backgrounds. The simulated signal is also represented. ging is employed.For this analysis, the preselection plays a crucial role,given the size of the MJ background. As an example,at D0, the preselection uses the following main require-ments. The events must have a well-reconstructed inter-action vertex and two or three jets with associated tracksto ensure efficient operation of the b -tagging algorithm.These jets must have p T >
20 GeV and | η | < E / T must begreater than 40 GeV, with large significance, i.e. with E / T values that are less likely to have originated fromfluctuations in jet energies, and the scalar sum of thetwo leading jet p T must be greater than 80 GeV.The dominant signal topology is a pair of b jets recoil-ing against the E / T due to the neutrinos from Z decaywith the direction of the E / T at large angles relative toboth jet directions. Conversely, in the case of events fromMJ background with fluctuations in jet energy measure-ments, the E / T tends to be aligned with a mismeasuredjet. An alternate estimate of E / T can be obtained fromthe missing p T calculated from the reconstructed chargedparticle tracks. This variable is less sensitive to jet en-ergy measurement fluctuations and, in signal events, isalso expected to point away from both jets, while in theMJ background its angular distribution is expected to bemore isotropic. A variable characterizing these featuresis used to further reject the MJ background. As for theother b ¯ b analyses, the physics backgrounds are taken intoaccount using Monte Carlo event generators.The preselected samples are subdivided into orthogo-nal subsamples based on the number and the quality ofthe b -tagged jets in the event. CDF has three taggingcategories ( SS , SJ , S ) for the analyzed two and three jetsamples, where S represents a jet identified by a recon-structed secondary vertex, and J represents a jet identi-fied by the presence of tracks not pointing to the maininteraction vertex. D0 has two tagging categories for itstwo jet sample ( T T and LL or T ). To suppress the4remaining backgrounds multivariate discriminant tech-niques are applied to each subsample as in Fig. 6. E. CDF and D0 results on H → b ¯ b searches Here we present the individual results from each col-laboration. The results are extracted using the MVAdiscriminant distributions from each subchannel of thesethree H → b ¯ b analyses, and then combined. TheCDF+D0 combination is discussed in Sec. IX. Here wepresent the results of each collaboration, while the limitsfor the three different search topologies are given in Ta-ble IV. The statistical techniques used are described inSec. IX and allow for the extraction of the limit on thesignal cross section normalized to the SM expectation, or,in case of excess, determine the p -value of the backgroundfluctuation. At very low mass, when combining the three H → b ¯ b topologies, CDF and D0 exclude at 95% C.L.Higgs bosons with masses smaller than 96 and 102 GeV,respectively. However, in the results from both collabo-rations an observed limit above the background-only ex-pectation is obtained for the ≃ m H = 125GeV are 1.8 or 4.2 and 2.3 or 3.2 times the SM expec-tation, for the CDF [123] and D0 [124] searches, respec-tively. To quantify the excess, the local p values are cal-culated and found to be minimal for a Higgs boson massof 135 GeV at CDF and D0, where the local significanceof these deviations with respect to the background-onlyhypothesis is 2.7 σ (1.7 σ ), which themselves correspondto 2.5 σ (1.5 σ ) global significances after applying look-elsewhere factors (cf. Sec. IX). These two mass valuesare compatible given the resolution of the dijet mass inthese final states. F. Searches in τ h final states Higgs boson searches using tau leptons decayinghadronically ( τ h ) complement those using electrons andmuons. CDF performs a generic analysis searching forHiggs bosons decaying to τ lepton pairs originating fromdirect gg → H production, associated W H or ZH pro-duction, and vector boson fusion production [125]. Afinal state consisting of one leptonic τ decay and onehadronic τ decay or two leptonic τ decays of differentflavors eµ is required. CDF hadronic τ identification isperformed using an MVA approach. The final discrimi-nant for setting limits is obtained combining the outputof four MVAs trained to separate a potential signal fromeach of the four primary backgrounds ( Z → τ τ , t ¯ t , multi-jet, and W +jet production). CDF also has an analysis ofevents that contain one or more reconstructed leptons ( ℓ = e or µ ) in addition to a τ lepton pair focusing on asso-ciated production where H → τ τ and additional leptonsare produced in the decay of the W or Z boson [126].Events are separated into five separate analysis channels ( ℓℓℓ , eµτ h , ℓℓτ h , ℓτ h τ h , and ℓℓℓℓ ). The four lepton cate-gory includes τ h candidates. The final discriminants arelikelihoods based on outputs obtained from independentMVA trained against each of the primary backgrounds( Z +jets, t ¯ t , and dibosons).The D0 ℓτ h jj analyses also include direct gg → H pro-duction, associated W H or ZH production, and vectorboson fusion production [127, 128]. Decays of the Higgsboson to tau, W , and Z boson pairs are considered. A fi-nal state consisting of one leptonic τ decay, one hadronictau decay, and two jets is required. Both muonic andelectronic subchannels are considered. The outputs ofboosted decision trees are used as the final discriminant. G. Searches in H → γγ Both CDF [129] and D0 [130] searched for Higgs bosonsdecaying into diphoton pairs with the full statistics (10fb − ). The CDF analysis searches for a signal peak inthe diphoton invariant mass spectrum above the smoothbackground originating from QCD production in sev-eral detector based categories with different signal-to-background ratios. In the D0 analysis the contributionof jets misidentified as photons is reduced by combininginformation sensitive to differences in the energy deposi-tion from real or false photons in the tracker and in thecalorimeter in a neural network output (NN o ). The out-put of an MVA, rather than the diphoton invariant mass,is used as the final discriminating variable. The finalMVA takes as input variables the NN o , the transverse en-ergies of the leading two photons and the azimuthal open-ing angle between them, the diphoton invariant mass andtransverse momentum, and additional variables, bringingan improvement in sensitivity of ≈ H. Searches for t ¯ tH production The t ¯ tH production is interesting for the direct tH cou-pling it involves, however its cross section is too smallat the Tevatron to contribute strongly to the overallsearch sensitivity. CDF uses several nonoverlapping setsof events to search for the process t ¯ tH → t ¯ tb ¯ b . Eventswith a reconstructed lepton, large missing transverse en-ergy, and four, five, and six or more jets are further sub-divided into five b -tagging categories [131]. Neural net-work discriminants are used to set limits. Events withno reconstructed lepton [132] are separated into two cat-egories, one containing events with large missing trans-verse energy and five to nine reconstructed jets and an-other containing events with low missing transverse en-ergy and seven to ten reconstructed jets. A minimum oftwo b -tagged jets is also required and events with threeor more b tags are analyzed separately from those withexactly two tags. Neural network discriminants are usedto reject large MJ background contributions and separatepotential t ¯ tH signal events from t ¯ t background events.5 VI. TEVATRON HIGH MASS HIGGS BOSONSEARCHES
As the hypothesized source of electroweak symmetrybreaking, the Higgs boson has strong coupling to bothmassive electroweak bosons. At Higgs boson massesabove 135 GeV the decay to a pair of W bosons is dom-inant, but even below the threshold to produce on-shell W bosons, the decay rate to one real and one virtual W boson is substantial. Both experiments pursue a strat-egy of searching for H → W + W − ( ∗ ) decay in final stateswith at least one charged lepton and from all produc-tion processes with substantial cross section. In addi-tion, searches for the subdominant decay H → ZZ areperformed. The high-mass search has similar sensitiv-ity to the individual searches for associated Higgs bosonproduction with a W or Z boson performed using the H → bb decay mode at a Higgs boson mass of 125 GeV.As proof of principle the experiments have observed allof the direct diboson production processes with pairs ofheavy gauge boson in final states that are topologicallysimilar to those used in the Higgs boson search. A. Diboson analysis:
W W , W Z and ZZ At the Tevatron, the diboson analyses in final stateswith charged leptons are performed using the same tech-niques used in the high mass Higgs boson search. The di-boson searches based on leptonic and semileptonic decaymodes allow for testing analysis techniques and develop-ing further understanding of the primary backgrounds tothe Higgs boson search. The direct SM production of the
W W [133, 134],
W Z [135], and ZZ [136] have been ob-served in leptonic decay modes with two charged leptons,three charged leptons and four charged leptons respec-tively. The diboson searches have also been performedwith larger datasets using the most modern lepton se-lections, providing measurements of the W W [137, 138]
W Z [139, 140] and ZZ [140, 141] cross sections. Thecombined production of W W and
W Z boson pairs hasbeen observed in events with one charged lepton andjets [142] and such an approach has been applied to per-form another high-mass Higgs boson search [143]. Thecombined production of all pairings of massive vectorbosons has been observed in events with one vector bosondecaying leptonically and the other hadronically [144].
B. Analysis Topologies
The analysis requirements typically require a triggeredlepton with p T >
20 GeV, possibly additional leptonswith lower thresholds, and significant E / T that is notaligned along the direction of other physics objects in theevents. The events are then categorized into a large num-ber of topologies that are consistent with various Higgsboson production and decay modes. These topologies are R(ll) D E ve n t s / . R(ll) D E ve n t s / . W+jets g WttWZZZDYWW 10 · HWW Data
CDF Run II Preliminary
OS 0 Jets = 160 GeV/c H M -1 L = 9.7 fb (cid:242)
FIG. 7: The angle between the lepton candidates in threedimensions: The distribution of this angle for backgroundand a H → W W signal with m h = 165 GeV are compared characterized by the number of charged leptons, whetherthe leptons are the same or opposite charge, and thenumber of jets. Each topology involves a limited set ofdominant signals and backgrounds allowing for optimaldiscrimination, and is therefore analyzed separately. Themost sensitive analysis topology involving zero jets andleptonic H → W + W − is described in detail below, whilethe subdominant modes are briefly discussed afterward.i) ggH → W + W − → ℓ + νℓ − ¯ ν + n j jet [145, 146]. When n j = 0, the signature is two opposite sign leptons, E / T ,and no observed jets. The signal in this final state isalmost 100% produced by the ggH process. The domi-nant background is from SM direct W W production withminor contributions from Drell-Yan production, the
W Z and ZZ diboson processes where one or more chargedleptons are not detected, and W +jets or W + γ wherea jet is misidentified as a lepton or the γ converts toan electron-positron pair, only one of which is detected.The strongest discriminant is the opening angle betweenthe leptons in two dimensions (2D) ∆ φ or three dimen-sions (3D) ∆ R (Fig. 7), due to the spin correlation be-tween the two spin one W bosons when decaying fromthe scalar Higgs boson. The collinear topology of thecharged leptons also results in a low dilepton invariantmass while Drell Yan background peaks at the Z massand other backgrounds at large mass. The neutrinos arealso collinear leading to larger E / T . Matrix element prob-abilities are effective discriminants because in the zerojet topology the final state of either Higgs boson or SMdirect W W production is well described by a leading or-der matrix element. The transverse mass of the Higgsboson can be well reconstructed since the neutrinos arealso collinear. The D0 experiment further subdividesthis mode by lepton flavor. CDF subdivides this analysisinto modes with two well-identified leptons, or one well-identified lepton plus and an isolated track, and analyzesevents with low dilepton invariant masses as a separate6 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 E ve n t s / . E ve n t s / . Wj g WttWZZZDYWW 10 · HWW Data
NN Output
CDF Run II
OS 0 Jets, High S/B = 165 GeV/c H M -1 L = 9.7 fb (cid:242)
FIG. 8: NN discriminant for ggH → W W → ℓνℓν + 0 jetat CDF. The distribution for background and a H → W W signal with m H = 165 GeV (multiplied by 10) are compared. category.When n j = 1 ,
2, the signature is two opposite signleptons, E / T , and observed jets. These topologies havesubstantial contributions from V H or VBF where the jetsare observed from either one of the vector boson decays orfinal state quarks respectively and additional backgroundfrom top pair production.ii)
V H → V W + W − → ℓ + Xℓ + ν +X [147, 148]. Asso-ciated production events can result in events with eithersame sign leptons or trileptons when the associatedvector boson decays to charged leptons. The backgroundincludes W +jets with a misidentified lepton in thesame sign mode and SM direct W Z production in thetrilepton case.iii) ( ggH, V H,
VBF) → H → W W → ℓν + ≥ W bosons decays leptonicallyand the other W boson decays to two quark jets [149].The dominant backgrounds are from W +jets and multi-jet background where a jet is misidentified as a lepton.iv) Other decay modes: The CDF and D0 experimentsalso consider modes where one W boson decays to a τ lepton which decays hadronically [150].Finally a search for the Higgs boson is performed in the H → ZZ mode where both Z bosons decay to chargedleptons [151]. The only significant background in thismode is SM direct ZZ production. The Higgs bosoncan be detected by looking for a narrow resonance inthe four lepton invariant mass distribution. D0 includesacceptance for H → ZZ in other cases where less thanfour charged leptons are found in the above searches. C. CDF and D0 results at high Higgs boson mass
All these search channels select a total of approxi-mately 75 Higgs boson events per detector for a Higgs bo-son mass of 165 GeV. More than half of these events are distinguished from the background with a good signal-to-background ratio using MVA discriminants as illustratedin Fig. 8. Examining the ∆ R distribution (Fig. 7) showsthat the largest discriminating power comes from the spincorrelation variable, due to the unique scalar nature ofthe Higgs boson.No significant excess is seen in any of the high-massHiggs boson search modes and limits are thus extracted,taking into account systematic uncertainties. The the-ory uncertainties become larger in events with more jetssince, for instance, in the NNLO calculation, events withtwo additional jets are calculated only at NLO accuracy.These uncertainties are addressed following the treat-ment by [152, 153] and included in the limit extraction.The expected limits with respect to the SM expec-tation for each experiment using the combination ofall high-mass Higgs boson search topologies at m H =165 GeV are 0.69 and 0.72 times the SM Higgs bosoncross section for the CDF [154] and D0 [155] searches, re-spectively. The experiments each achieve expected sen-sitivity within a factor of approximately 1.5 of the SMcross section, in the mass range m H = 140 −
185 GeV.The CDF (D0) analysis excludes the SM Higgs boson inthe mass range m H = 148 −
175 GeV (157 −
172 GeV).Additionally these searches provide strong sensitivity tothe production of a Higgs boson at lower masses. Theexpected or observed limits at m H = 125 GeV are 3.1 or3.0 and 3.6 or 4.6 times the SM Higgs boson cross sectionfor the CDF and D0 searches, respectively. The limitsfor the different search topologies at m H = 125 GeV aregiven in Table IV. VII. LHC SEARCHES IN BOSONIC HIGGSBOSON DECAYS
The Higgs boson searches at the LHC are performedby decay mode. In this section we discuss each searchseparately, while the combination of all search resultsfor each experiment is discussed in Sec. IX. Previoussearches at LEP, Tevatron, and the LHC, in addition toindirect constraints indicated that the SM Higgs bosonhad a low mass between approximately 115 and 130 GeVwith the region around 125 GeV being of highest interest.As of 4 July 2012 the LHC experiments had analyzedsearches sensitive to this mass range and higher massesusing datasets of approximately 5 fb − at 7 TeV and 5.8(5.3) fb − for ATLAS (CMS) at 8 TeV. A. LHC diboson physics
The key modes for observing the Higgs boson at lowmass are the fully reconstructed γγ and ZZ decay modes.To explore the role of the Higgs boson in electroweaksymmetry breaking the W W decay mode is also crucial.Understanding the non-resonant continuum productionof these final states is important to control backgrounds7in the Higgs boson searches. The LHC experiments haveobserved γγ production in 7 TeV pp collisions [156]. W W [157–162] and ZZ [163–166] production has beenobserved in both 7 and 8 TeV collisions and their crosssections measured. These measurements typically useidentical selections and techniques to those of the Higgssearches. Large samples have been collected in each de-cay mode. With future data samples the W W and ZZ modes can be used to study the electroweak symmetrybreaking related phenomena of longitudinal vector bosonscattering. The contribution of Higgs boson exchange tothis process should limit the otherwise divergent behaviorof this process at high energy. B. Searches in H → γγ The LHC experiments have the ability to reconstructa Higgs boson in the two photon decay mode [167–170].The detectors are designed with the excellent energy andposition resolution necessary to accurately reconstructthe invariant mass of two photon events. Excellent massresolution is critical since backgrounds from multijet,multijet+photon, and photon+photon events are large.This decay has a small branching ratio, since the Higgsboson can decay only to photons through a loop diagraminvolving massive charged particles. The large inclusiveproduction cross section at lower masses makes this a vi-able mode for a Higgs boson observation at low mass. Inaddition, observation of this mode would rule out spin 1for the observed object.Data are collected with diphoton triggers. Energy andisolation requirements are made on the photons to reducebackgrounds. Converted photons are also used and pro-vide good energy and position resolution when electronpairs can be reconstructed by the trackers. The ATLASexperiment reduces background by using the longitudinalsegmentation of the calorimeters to select photons thatpoint to the hard interaction vertex.Both experiments apply a separate selection for eventswith two forward jets to focus on a vector boson fusionlike topology which has excellent sensitivity achievinga signal to background that is an order of magnitudegreater than in topologies without forward jets typicalof gluon fusion production. In the ATLAS experiment,to optimize sensitivity, events are further divided bywhether photons are converted or not and by which pseu-dorapidity region of the detector they are reconstructedin, since these classes result in different diphoton massresolution and as a result different signal to background.Finally, events are classified by the transverse momen-tum of the diphoton system since backgrounds are sub-stantially reduced at high values. In the CMS experimentthe output of a dedicated photon identification BDT, thetransverse momentums of the photons, the opening an-gle between the photons, the pseudorapidity of the pho-tons, and the estimated mass resolutions of the diphotonsystem are used to classify events by expected signal-
100 110 120 130 140 150 160 E v en t s / G e V ATLAS ggfi H DataSig+Bkg FitBkg (4th order polynomial) -1 Ldt=4.8fb (cid:242) =7 TeV, s -1 Ldt=5.9fb (cid:242) =8 TeV, s (a) =126.5 GeV) H (m
100 110 120 130 140 150 160 E v en t s - B k g -200-1000100200 (b)
100 110 120 130 140 150 160 w e i gh t s / G e V S Data S/B WeightedSig+Bkg FitBkg (4th order polynomial) =126.5 GeV) H (m (c) [GeV] gg m100 110 120 130 140 150 160 w e i gh t s - B k g S -8-4048 (d) FIG. 9: Invariant mass of diphoton events in the ATLASexperiment. The results are presented with and without eventweighting by an expected signal-to-background ratio. to-background ratios using a BDT multivariate discrimi-nate. An investigation of the categories formed by select-ing on the BDT output indicates that divisions based onthe transverse momentum of the diphoton system, the de-tector pseudorapidity region of the photons, and whetherthe photons are converted or unconverted largely deter-mines the classification of events.The sensitivities of the individual categories are takeninto account when calculating the overall sensitivity ofthe analysis and when determining exclusions or signalsignificances. The collaborations weight events by ex-pected signal-to-background ratio when displaying thediphoton mass distribution to give a visual representa-tion of the benefit of this classification.Using the two photon invariant mass distribution tosearch for the Higgs boson, regions at both low massand higher mass have been excluded by both experimentsleaving only a narrow region of mass unexcluded. TheATLAS experiment excludes the regions 112-122.5 and132-143 GeV. This exclusion extends the lower exclusionbound of the LHC searches below the upper exclusionbound from the LEP searches, and when combined withother searches excludes the entire mass range below 600GeV except for the narrow allowed region around 125GeV.These analyses have a strong sensitivity to the produc-tion of a low-mass Higgs boson of around 125 GeV with8 (GeV) γγ m
110 120 130 140 150 S / ( S + B ) W e i gh t ed E v en t s / . G e V DataS+B FitB Fit Component σ ± σ ± -1 = 8 TeV, L = 5.3 fbs -1 = 7 TeV, L = 5.1 fbsCMS (GeV) γγ m
120 130 E v en t s / . G e V Unweighted
FIG. 10: Invariant mass of diphoton events in the CMSexperiment with events weighted by an expected signal-to-background ratio and the unweighted distribution shown asan inset. the expectation of observing approximately 200 eventsper experiment at that mass. In that region both ex-periments see a significant excess of events. The ATLASdiphoton invariant mass distributions both weighted bysignal-to-background ratio and unweighted are shown inFig. 9. The corresponding CMS diphoton invariant massdistributions are shown in Fig. 10.The ATLAS experiment sees a 4 . σ excess of eventscompatible with a narrow resonance of mass 126.5 GeVwith a signal strength of 1 . ± . . σ excess of eventscompatible with a narrow resonance of mass 125 GeVwith a signal strength of 1 . ± . C. Searches in H → ZZ → ℓ + ℓ − ℓ + ℓ − Unique to the LHC experiments is the ability to ob-serve the Higgs boson over a large range of massesthrough inclusive production and H → ZZ → ℓ + ℓ − ℓ + ℓ − decay [171–174]. In the CMS search the charged lep-tons from Z decay considered at 7 TeV include the τ lep-tons [175]. The LHC experiments are designed to providelarge angular coverage for lepton identification in orderto detect events in this mode at an adequate rate. Be-cause of the excellent lepton momentum resolution of theexperiments, the Higgs boson mass can be reconstructedwith sufficient precision that background rates are verylow and event counts on order of 10 events are sufficientfor discovery. With the assumption of SM production rate the coupling to the Z boson can be measured, and,by comparison with the W W mode, the ratio of W and Z couplings can be measured. Finally, with larger datasamples than reported here, the spin and parity of apossible Higgs boson can be determined solely from thismode using an angular analysis of the decay products.Data are collected using single and dilepton (CMS)triggers. Transverse momentum requirements are madeon the leptons to reduce backgrounds. Advanced tech-niques such as multivariate lepton identification are ap-plied to maximize lepton finding efficiency. The CMSexperiment improves mass resolution by using an algo-rithm designed to detect and recover the momentum offinal state photons radiated by the leptons. ATLAS in-corporates a similar technique as part of its electron mo-mentum fit. In ATLAS one opposite charge same flavorpair of leptons is required to be consistent with the Z bo-son mass. After these requirements the most significantbackground is direct SM ZZ production. A Higgs bosonsignal can be distinguished from the background by look-ing for a narrow resonance in the four lepton invariantmass distribution. For a Higgs boson mass of 125 GeVthe ATLAS and CMS experiments expect ≃
10 eventseach.The performance of the four lepton search includingselection efficiency estimates and the scale and resolu-tion of the four lepton invariant mass is tested by bothexperiments by searching for the four lepton final stateproduced by a Z boson where one initial decay lepton ra-diates a photon which converts to a lepton and antileptonpair Z → ℓ . With looser selection on lepton transversemomentum and the dilepton masses of same flavor op-posite sign pairs this mode can be detected with sub-stantially higher statistical precision than a Higgs bosonat lower mass. Both experiments perform this analysiswithin the framework of the Higgs boson search anal-ysis and observe this decay with the expected perfor-mance [172, 174, 176].In both experiments the dominant SM ZZ backgroundis estimated from simulation normalized to NLO crosssection predictions. In ATLAS the backgrond from ℓ + ℓ − +X events, which is dominated by Z + b ¯ b and t ¯ t events, is estimated by measuring the normalization ofthese backgrounds using selections designed to select non-prompt muons or nonisolated electrons, which are morelikely to originate from b jets, and using a transfer factorfrom simulation to extrapolate the background predic-tion to the signal region. In CMS the backgrounds from Z +X events and t ¯ t are estimated in a control region withone same flavor opposite charge dilepton pair with an in-variant mass consistent with the Z boson and additionalobjects. Using a subset of events from this region withone additional identified lepton the probability for an ob-ject to be falsely identified as a charged lepton from Z decay is measured. That misidentification rate is appliedto determine the number of events with a Z boson andtwo additional lepton candidates that are not from Z bo-son decay in the signal region.9 [GeV] m100 150 200 250 E v en t s / G e V -1 Ldt = 4.8 fb (cid:242) = 7 TeV: s -1 Ldt = 5.8 fb (cid:242) = 8 TeV: s 4l fi (*) ZZ fi H Data (*)
Background ZZ tBackground Z+jets, t=125 GeV) H Signal (mSyst.Unc.
ATLAS
FIG. 11: The four lepton invariant mass distribution fromthe ATLAS experiment. The data are displayed as point andthe background expectation as a histogram. Several SM Higgsboson signal contributions are included for different hypothet-ical Higgs boson masses. Background Z +jets and t¯t bottom,background ZZ middle and Higgs boson signal top. Both experiments use the four lepton invariant massdistribution to search for a Higgs boson. CMS furtheruses angular information based on the expected scalarspin zero and parity even nature of the Higgs boson ina matrix element likelihood analysis(MELA). Based onthe searches large regions at high mass are excluded.The ATLAS experiment excludes the regions 131-162 and170-460 GeV, while the CMS experiment excludes theregions 131-162 and 172-525 GeV. These are the largestexclusions from a single analysis channel, failing to ex-clude only the mass region where the
W W branchingratio dominates just above the on-shell
W W productionthreshold, and in the low-mass region.These analyses have strong sensitivity to the produc-tion of a low-mass Higgs boson of roughly 125 GeV,as can be seen from the four lepton mass distributionfrom ATLAS, shown in Fig. 11, and from the invariantmass distributions of the four lepton candidates versusthe MELA discriminant from CMS shown in Fig. 12.Both experiments see a significant excess of events.The ATLAS experiment sees a 3.4 σ excess of events com-patible with a narrow resonance of mass 125 GeV witha signal strength of 1.3 times the SM expectation. TheCMS experiment sees a 3.2 σ excess of events compatiblewith a narrow resonance of mass 125.6 GeV with a signalstrength of approximately 0.7 times the SM expectation.The evidence presented by both experiments of a narrowresonance with decays to ZZ indicates that a new bosonhas been observed, and the use of angular information toenhance the signal in the CMS case weakly favors spinzero and parity even as quantum numbers for the new bo-son though no definitive measurement of these propertiesis yet possible. Further characterization of this excess is FIG. 12: A two-dimensional plot of four lepton invariant massversus matrix element likelihood from the CMS experiment.Data are shown with event by event mass uncertainties whilethe expectation of a 125 GeV SM Higgs boson is superimposedas a temperature plot. The central region around 125 GeV ishighest in probability. given next.
D. Searches in H → W + W − → ℓ + νℓ − ¯ ν The LHC experiments search for inclusive Higgs bo-son production with the decay H → W + W − → ℓ + νℓ − ¯ ν [177–180]. The decay is not fully reconstructedbecause of the neutrinos in the final state. However,the observation of collinear charged leptons in this de-cay mode is a distinct signature for the decay of a scalarparticle. Observation of the Higgs boson in this decaymode also excludes spin one as a potential spin state.Also, using the production rate from theory, the modecan be used to determine the Higgs boson coupling tothe W boson or the ratio of W and Z couplings throughcomparison with the H → ZZ mode. In addition, bycomparing events with zero or one jet to events with twoforward jets, the gluon fusion and vector boson fusionproduction rates can be compared.The data for the searches are collected using single lep-ton triggers (ATLAS) and dilepton triggers (CMS). Re-quirements are made on the transverse momentum of thecharged lepton candidates, the magnitude of the miss-ing transverse energy and on the E / T direction, whichmust not be collinear with other physics objects in theevent. Also E / T measured using the calorimeter andtracker are compared to require compatibility and rejectevents with false E / T . Loose selection on the collinearityof the charged leptons is imposed to be consistent withthe decay of a spin zero object to a W boson pair withsubsequent leptonic decays.ATLAS considers only electron and muon events, sincesuch final state does not have significant Drell-Yan back-grounds and is substantially more sensitive. CMS dividesevents into same flavor, electron or muon, and differ-0ent flavor, electron and muon, subsamples. Both experi-ments divide the data into events with zero, one, or twoor more jets. The primary backgrounds are SM direct di-boson production, W + jets, Drell-Yan, single top, and tt . The tt background is dominant and the division byjet counting is designed to isolate the lower jet categorieswhich have smaller top contributions. In addition b -tagvetoes are applied including, at CMS, the vetoing of jetsunder a jet transverse momentum threshold. The twoexperiments apply additional selection criteria to furthertake advantage of the collinear nature of the charged lep-tons in Higgs boson decay to W bosons. In the two jetanalysis the jets are required to have a large rapidity dif-ference and large jet-jet invariant mass to be consistentwith the forward jets from vector boson fusion.Both experiments construct orthogonal control regionsto study background kinematics and normalize back-ground contributions. To study t ¯ t and tW backgroundsa region is constructed with no jet multiplicity selection(ATLAS), or requiring one b tagged jet above threshold(CMS). For zero jet events the background is normalizedfrom this region by extrapolating to the zero jet topologyusing top event kinematics from simulation. For eventswith jets the background is normalized in these regionsby applying b -tagging and b mistagging rates measuredin a t ¯ t dominated region to calculate the number of topevents that fail the b -tagging criteria in the data. In addi-tion the region with one b -tagged jet can be used to studythe performance of under-threshold b jets to cross-checkthe performance of the veto on b -tagged under-thresholdjets that is used in zero jet events used in CMS. A regionwith diepton invariant mass larger than 80 GeV (AT-LAS) or 100 GeV (CMS) is used to normalize the W W contribution. In ATLAS the contribution of Drell-Yanevents produced off the Z resonance is estimated usingsimulation after additional selection on E / T and the trans-verse momentum of the dilepton system to reduce thiscontribution. In CMS, this background is estimated bymeasuring the Z resonance rate, subtracting non-Drell-Yan contributions using µe events, and then extrapolat-ing the result to the off Z resonance mass range usingthe expected distribution of the dilepton mass from sim-ulation. W +jets backgrounds are estimated using a fullydata driven method relying on identifying a sample of W +jets events with a second lepton candidate passinga loose plus antilepton selection (ATLAS), or simply aloose selection (CMS), and applying false lepton identifi-cation rates measured from data.The ATLAS experiment uses the transverse mass ofthe Higgs boson to search for the signal, as shown inFig. 13, while CMS uses a BDT algorithm to distinguishsignal from background.These analyses extend the sensitivity of the searchesfor pairs of vector bosons to lower masses including sen-sitivity to masses as low as 125 GeV. The experimentsexclude the ranges greater than 137 GeV (ATLAS) and129-520 GeV (CMS). No upper limit is placed in the AT-LAS analysis as the analysis was optimized for low Higgs [GeV] T m
50 100 150 200 250 300 E v en t s / G e V Data stat) ¯ SM (sys WW g WZ/ZZ/Wt t Single Top Z+jets W+jets H [125 GeV]
ATLAS -1 Ldt = 5.8 fb (cid:242) = 8 TeV, s + 0/1 jets n e nm / nmn e fi (*) WW fi H FIG. 13: Distribution of the transverse mass in the zero-jetand one-jet
W W analyses of ATLAS, with both eµ and µe channels combined. The expected signal for m H = 125 GeVis shown stacked on top of the background prediction. boson mass values. Both experiments have the sensitivityto exclude a Higgs boson at smaller masses but observean excess around ≃
125 GeV. The ATLAS excess hasa local significance of 2.8 σ and corresponds to a signalstrength of 1 . ± . σ . E. Searches in H → W W, ZZ with decays of oneboson to quarks or neutrinos
The ATLAS and CMS experiments conduct inclu-sive searches for the Higgs boson in the decay modes H → W + W − → ℓ + νq ¯ q [181–183], H → ZZ → ℓ + ℓ − q ¯ q [184, 185] and H → ZZ → ℓ + ℓ − ν ¯ ν [186–188].Although these search modes have higher backgroundsthan those of the other decay modes to pairs of vectorbosons discussed earlier they have strong sensitivity forhigh-mass Higgs bosons where the vector bosons havehigh transverse momentum leading to significantly re-duced backgrounds.Data are collected on one and two lepton triggers forthe analyses with W and Z bosons, respectively. Thecharged leptons are required to pass transverse momen-tum and identification requirements. In the modes with Z → ℓ + ℓ − decays the invariant mass of the charged lep-tons are required to be consistent with the Z boson masswhile in the W → ℓν case the transverse mass formedfrom the E / T and the lepton momentum is required to beconsistent with the expected W boson transverse mass.In the cases where a Z boson decays to neutrinos the E / T is required to be large. The ATLAS experiment fur-ther divides this analysis into high and low Higgs bosonmass versions where the E / T is required to be larger inthe high-mass version. Finally, in the cases where the1vector bosons decay to quarks the jet-jet invariant massis required to be consistent with the expectation for a W or Z boson.The primary backgrounds are from SM diboson pro-duction, tt production, and QCD multijet productionwhere mismeasurement of a jet mimics one of the lep-tonic signatures. In all of these background processesthe candidate diboson pair is not expected to form amass resonance, and the individual vector bosons are notexpected to be boosted. The experiments apply criteriato exploit these characteristics including requirements onthe boost of individual vector bosons and the openingangle between vector boson decay products. After se-lection they search for the Higgs boson using either thefull mass reconstruction ( H → ZZ → ℓ + ℓ − q ¯ q ), trans-verse mass reconstruction ( H → ZZ → ℓ + ℓ − ν ¯ ν ) or fullmass reconstruction applying a W boson mass constraintin the H → W + W − → ℓ + νq ¯ q mode. The experimentssearch for a Higgs boson in the range 130-600 GeV, withvarying lower thresholds depending on the analysis, us-ing 4.7-5.0 fb − of 7 TeV collision data per experiment,while the CMS experiment additionally includes 5.1 fb − of 8 TeV collision data in the H → W + W − → ℓ + νq ¯ q and H → ZZ → ℓ + ℓ − ν ¯ ν modes. The combined resultsprovide substantial constraints on the mass of a high-mass Higgs boson excluding masses from 230 to 600 GeV.The exclusion is dominated by the H → ZZ → ℓ + ℓ − ν ¯ ν mode although extended in the lower mass range by the H → W + W − → ℓ + νq ¯ q search. VIII. LHC SEARCHES IN FERMIONIC HIGGSBOSON DECAYSA. Searches in H → τ + τ − The LHC experiments search for the SM Higgs bosonin the tau lepton pair decay mode [189–191]. The Higgsboson to τ lepton pair branching ratio is 8% to 1.5%in the Higgs boson mass range of 115-150 GeV, and τ lepton signals are distinct enough to make this a viablesearch mode for all Higgs boson production processes.However, the production mode with strongest sensitivityis the vector boson fusion production mode with two as-sociated forward jets. This mode is of high interest sinceobservation of Higgs boson production via vector bosonfusion gives direct information on how the Higgs bosoninteracts with high-energy longitudinal vector bosons. Inaddition it will allow the measurement of the coupling ofthe τ lepton to the Higgs boson which should be the firstmode to establish a clear signal in a fermionic decay atthe LHC and is the only accessible leptonic coupling inthe hadron collider environment, prior to the planned lu-minosity upgrades. The τ τ invariant mass obtained inthe CMS vector boson analysis is shown in Fig. 14The experiments search for events with zero, one, ortwo associated jets and oppositely charged τ lepton pairsin the following τ τ decay final states: ee (ATLAS), µµ , FIG. 14: Distribution of the τ τ invariant mass in the com-bined 7 and 8 TeV data sets for the VBF category of theCMS τ + τ − analysis. The expected signal for m H = 125 GeVis shown stacked on top of the background prediction. eτ h , µτ h , and τ h τ h (ATLAS), where e indicates τ → eνν , µ indicates τ → µνν , and τ h where τ h indicates the τ lepton decayed to hadrons and a τ neutrino. Data arecollected on triggers that require one or two charged lep-tons and in the ATLAS experiment one trigger requirestwo high- p T τ h candidates ( τ h τ h ).The primary backgrounds are Drell-Yan production, W +jets production, where one jet is misidentified as a τ lepton, and tt production. To enhance the presence ofa possible signal relative to backgrounds, the leptons orsum of hadronic decay products are required to have sub-stantial transverse energy and the E / T from undetectedneutrinos is required to be collinear with the direction ofthe dilepton system. In events with two jets, the two jetsare required to have high momentum and be containedwithin the forward calorimeters. There also must be alarge rapidity gap between the two jets consistent withthe vector boson fusion hypothesis. In events with onejet, the p T of the jet is required to be high to enhance theHiggs boson candidate boost, which improves separationof the Higgs boson signal from backgrounds and allowsfor a more precise estimate of the Higgs boson mass.The experiments search for the presence of a Higgsboson using methods designed to more fully reconstructthe mass of the Higgs boson by including the E / T in thecalculation and taking advantage of the boosted config-urations. The ATLAS (CMS) experiments search for aHiggs boson in the range 110-150 GeV (110-145 GeV) us-ing 4.7-4.9 fb − of 7 TeV collision data per experiment,while the CMS experiment additionally includes 5.1 fb − of 8 TeV collision data. The ATLAS experiment reachessensitivity to set limits in the range 3-11 times the ex-pected SM production rate, while the CMS experimentreaches sensitivities of 1.3-2.4 times the SM rate. At a2mass of 125 GeV the CMS experiment sets a limit onthe Higgs boson production cross section of a SM Higgsboson of 1.1 times the expected SM rate. The sensitivityof this analysis is sufficient to achieve evidence for thisdecay mode using tens of fb − . B. Searches in H → b ¯ b The LHC experiments conduct searches for associatedproduction of Higgs bosons with a W or a Z bosonwith subsequent decay of the Higgs boson to a pair of b -quarks [192–194]. The Higgs boson to b -quark pairsbranching ratio is 70% to 40% in the mass range of 115-135 GeV. However, the b -quark decay signature is notdistinct enough to extract the signal from the backgroundand the leptonic decay signatures of the massive vectorbosons produced in associated production are also nec-essary to search for the Higgs boson in this decay mode.This search is divided into three subsearches by the de-cay mode of the massive vector boson. The experimentssearch for W H → ℓνb ¯ b , ZH → ℓ + ℓ − b ¯ b , and ZH → E / T b ¯ b where the Z decays to neutrinos, “observed” as E / T . Acharged lepton ( ℓ ) refers to electrons and muons of bothelectric charges. Although this set of production and de-cay processes is less sensitive than those of many otherHiggs boson search modes, it is important because it caneventually be used to measure the relative couplings ofthe Higgs boson to the W and Z bosons and uniquelymeasure its coupling to b quarks.Data are collected on triggers that require a singlecharged lepton for events with W decays, single leptons(ATLAS ee mode), or pairs of leptons for Z decays tocharged leptons, and either E / T (ATLAS) or E / T +jets(CMS) for events with Z decays to neutrinos.Backgrounds such as W +jets, single top and tt , di-bosons, and QCD multijet production with misidentifiedleptons are several orders of magnitude larger than thesignal. To reconstruct a possible signal the analyses makeseveral additional requirements. Leptons must be fullyreconstructed by both the tracking system and dedicatedlepton identification systems and pass minimum trans-verse momentum thresholds. Missing transverse momen-tum must not be collinear with the jets. In addition, theATLAS experiment requires that E / T reconstructed usingcalorimeter and tracking-based algorithms is consistent,while the CMS experiment uses particle flow-based ob-jects to take advantage of all detector subsystems. Twojets must be identified as b jets, as expected from the b -quarks from the Higgs boson decay.Finally in the ZH → E / T b ¯ b mode the E / T is requiredto be very large. The ATLAS experiment separates theanalysis into separate channels for different ranges of vec-tor boson p T , while the CMS experiment selects eventsbased on the transverse momenta of both the vector andHiggs boson candidates.The ATLAS experiment uses the invariant mass of thetwo b jets to search for Higgs boson candidates within each channel, and combines the results into a singlesearch. The CMS experiment uses variables associatedwith the above quantities and the dijet invariant mass ofthe Higgs boson as inputs to multivariate discriminantsused to distinguish signal from background. The ATLAS(CMS) experiment search for a Higgs boson in the range110-130 GeV (100-135 GeV) using 4.7-5.0 fb − of 7 TeVcollision data per experiment, while the CMS experimentadditionally includes 5.1 fb − of 8 TeV collision data.The CMS analyses have sensitivity to set limits on theorder of 1 to 5 times the expected SM Higgs boson crosssection, depending on the mass, achieving a sensitivity of1.6 at a Higgs boson mass of 125 GeV. The sensitivity ofthis analysis is be sufficient to achieve evidence for thisdecay mode using tens of fb − . C. Searches for t ¯ tH production The LHC experiments search for associated productionof the Higgs boson with t ¯ t where the Higgs boson is ra-diated from one of the top quarks. The strong couplingbetween the Higgs boson and the top quark increasesthe probability of such radiation. This process providesa direct measurement of the top-quark Yukawa coupling,which is expected to be one indicating maximal coupling.Statistically it will not be as significant as measuring thecoupling in gluon fusion events, which are dominated bythe top loop diagram, but it will not suffer from theo-retical uncertainties associated with understanding thegluon fusion loop process. The ATLAS and CMS exper-iments have performed an analysis with the full 7 TeVdata set [195, 196] in the final states with two leptons(CMS only) and with lepton plus jets decay modes of thetop quarks with Higgs boson decay to b ¯ b . Events with4-6 (2-6) jets and 0-4 (2-4) b -tagged jets are consideredat ATLAS (CMS). The ATLAS experiment uses the m bb and H T (total energy) distributions to set limits, while theCMS experiment uses an MVA discriminant. The CMSexperiment has sensitivity to set limits on the order of3 to 9 times the expected SM Higgs boson cross sectionachieving a sensitivity of 4.6 at a Higgs boson mass of125 GeV. With a data set on the order of hundreds offb − it is be possible to directly measure the top quarkYukawa coupling in this channel. IX. ATLAS, CMS, AND TEVATRON RESULTSA. Limits and combination methods
At the LHC and Tevatron limits are calculated usingthe modified frequentist CL s approach. At the Teva-tron, a Bayesian technique is also used. The techniqueshave been shown to produce similar results at the level ofabout 5%. To facilitate comparisons with the SM and toaccommodate analyses with different degrees of sensitiv-ity and acceptance for more than one signal production3mechanism, the limits are divided by the SM Higgs bo-son production cross section, as a function of Higgs bosonmass, for test masses for which the experiments have per-formed dedicated searches in different channels. A valueof the combined limit ratio R which is less than or equalto 1 indicates that that particular Higgs boson mass isexcluded at the 95% C.L. Expected limits are calculatedboth for the background only hypothesis ( B ), for whichonly SM background contributions are present in the se-lected data samples, and for the signal-plus-backgroundhypothesis. The signal-plus-background hypothesis iscalculated by also including the simulated signal contri-bution in the limit setting procedure. The limits are gen-erally determined using the MVA output distributions orthe invariant mass distributions, together with their asso-ciated uncertainties, as discriminating inputs to the limitsetting procedure.In the CL s approach, each hypothesis is tested by sim-ulating the outcome of multiple pseudoexperiments. Thedata are assumed to be drawn from a Poisson statisticalparent distribution, and each pseudo experiment resultis obtained by randomly generating pseudodata using aPoisson distribution for which the mean is taken from ei-ther the background-only or signal-plus-background hy-pothesis. To evaluate the statistical significance of eachresult a negative Poisson log-likelihood ratio (LLR) teststatistic is evaluated, and the outcomes are ordered interms of their contributing statistical significance. Thefrequency of each outcome is used to define the shape ofthe resulting LLR distributions at each mass point forboth the background-only and signal-plus-backgroundhypotheses.Systematic uncertainties in each hypothesis are ac-counted for by nuisance parameters which are assignedan a prior probability distribution. These parametersrefer to uncertainties in the expected background contri-butions and, in the case of the signal-plus-backgroundhypothesis, also uncertainties on the simulated signalcontribution. The nuisance parameters are Gaussianand randomly assigned within the parent distributionfor each pseudo experiment. Correlations between theuncertainties are taken into account. To minimize theimpact of the nuisance parameters the profile likelihooddistribution is maximized over the nuisance parameterswithin each pseudo experiment, once for the background-only and once for the signal-plus-background hypotheses.Each background is allowed to vary within its uncertain-ties by varying the nuisance parameters in the fittingprocedure, while the fit is constrained to lie within theuncertainties.The expected limits are calculated with respect tothe median of the background-only LLR distribution,whereas the observed limits are quoted with respect tothe single LLR value of the actual measurement. Thedistribution of expected limits can also be analyzed tounderstand 1 σ and 2 σ deviations from median.This framework can also produce statistical resultsquantifying the expectation for and properties of a sig- nal. Given the SM expectation for signal contributions,the expected p -value or probability for backgrounds tofluctuate to the statistical significance of the expectedsignal can be computed. Similarly, given an excess in thedata, the observed p -value can be computed. Finally, inthis technique the SM Higgs signal cross section is mul-tiplied by an arbitrary factor that is fit for the likelihoodminimization allowing for a measurement of the observedcross section. B. ATLAS and CMS results
The ATLAS and CMS experiments analyze their datausing the statistical techniques described previously.Each analysis channel is analyzed separately and withineach experiment the Higgs boson search results are com-bined [197–201]. Table III summarizes for ATLAS andCMS, the integrated luminosities, the Higgs boson massranges over which the searches are performed, and refer-ences to further details for each analysis. Also given areexpected and observed exclusion ranges for SM Higgs bo-son production at 95% C.L. or for those channels whichdo not have sensitivity to limit the SM rate of Higgs bo-son production at any mass, the expected and observedlimits on cross section for the SM Higgs boson with mass m H = 125 expressed as a multiplicative factor times thepredicted SM rate. Depending on the mass of a hypothet-ical Higgs boson, the LHC experiments have the sensitiv-ity to discover the Higgs boson in individual productionand decay channels. At some masses it is possible to haveobservations of the Higgs boson in several channels. Thecombination of these results allows the experiments toachieve larger exclusion ranges over masses where no ev-idence for a Higgs boson signal is seen, earlier discoveriesin mass ranges where several analysis channels have sen-sitivity, and comparison among channels to demonstratethe consistency of a possible signal with a Higgs bosonhypothesis. The LHC experiments are prepared for, buthave not yet produced a joint combination of these re-sults. For the first discovery the combination of resultsfrom multiple experiments is not preferred since simul-taneous observation constitutes both an observation andan independent confirmation of the result.Over a large region of masses the LHC experimentsobserve no evidence for a Higgs boson. The LHC datashow a consistent picture with a high-mass SM Higgsboson typically excluded by multiple channels. At highmass the ATLAS experiment excludes the production ofa SM Higgs boson with masses from 131 to 559 GeVat 95% C.L. and the CMS experiment excludes a regionfrom 128 to 600 GeV at 95% C.L., where 600 GeV is thelimit of the search range. At low mass the ATLAS exper-iment excludes the production of a SM Higgs boson withmasses from 111 to 122 GeV at 95% C.L., while the CMSexperiment excludes the region from 110 to 122.5 GeV.4 TABLE II: The most significant excesses seen in ATLAS and CMS results and the combined local and global significances orP values. Topology ATLAS Significance and Mass CMS Significance and Mass H → W W → ℓνℓν σ σ H → ZZ → ℓ σ σ H → γγ σ σ σ σ R ) assuming m H = 125 are given. For analyses with SM sensitivity expected and excluded rangesof mass are given.ATLAS Channels Luminosity m H range Expected exclusion Observed exclusion Reference(7+8 TeV, fb − ) (GeV) (Range in GeV, R ) (Range in GeV, R ) ttH → ttbb V H → V b ¯ b R =4.0 R =4.6 [192] H → τ + τ − R =3.3 R =3.4 [189] H → γγ H → W W → ℓνℓν > > H → W W → ℓν H → ZZ → ℓ H → ZZ → ℓ H → ZZ → ℓ ν m H range Expected exclusion Observed exclusion Reference(7+8 TeV, fb − ) (GeV) (range in GeV, R ) (range in GeV, R ) ttH → ttbb R =4.6 R =3.8 [196] V H → V b ¯ b R =1.6 R =2.1 [193, 194] H → τ + τ − R =1.28 R =1.06 [190, 191] H → γγ H → W W → ℓνℓν H → W W → ℓν H → ZZ → ℓ H → ZZ → ℓ τ H → ZZ → ℓ H → ZZ → ℓ ν The combined ATLAS and CMS limits are presented inFigs. 15 and 16, respectively.At low masses the experiments have the sensitivity toexclude or observe the Higgs boson. The sensitivities forobservation of a signal are quantified as an expected p -value for the background to fluctuate to a signal as largeas the median expectation for a SM Higgs boson. Thecombined expected p -value at m H = 125 GeV is 4 . σ for the ATLAS experiment and 5 . σ for the CMS exper-iment. In the region around 125 GeV both experimentsobserve an excess of events in multiple search channels.The experiments evaluate the p values for each channelseparately and for the entire combination and comparethose values with the expected background-only p val-ues given a SM Higgs boson as a function of mass (seeFig. 17,18). Information quantifying the most signifi- cant excesses in the individual search channels was givenpreviously in the sections describing the different LHCHiggs boson searches and is summarized along with themost significant combined excess from each experimentin Table II. Both experiments observe a Higgs boson sig-nal with local significances above the evidence level of3 σ in the ZZ and γγ decay modes and combined sig-nificances of 5 . σ at m H = 126 GeV for the ATLASexperiment and 5 . σ at m H = 125 . . σ . The simultaneous observation of anew particle with mass of approximately 125 GeV con-stitutes a definitive discovery. The decay modes in whichthe particle is strongly observed also indicate that the5 [GeV] H m100 200 300 400 500 600 S M s / s % C L L i m i t on -1 Obs. Exp. s – s – -1 Ldt = 5.8-5.9 fb (cid:242) = 8 TeV: s -1 Ldt = 4.6-4.8 fb (cid:242) = 7 TeV: s
ATLAS
Preliminary
CLs Limits
FIG. 15: The ATLAS experiments combined upper limit as afunction of the Higgs boson mass between 100 and 600 GeVSolid black: observed limit/SM; dashed black: median ex-pected limit/SM in the background-only hypothesis: col-ored bands: ± , σ distributions around the median expectedlimit. Higgs boson mass (GeV)
100 200 300 400 500 H S M s / s % C L li m i t on -1 ObservedExpected (68%)Expected (95%)ObservedExpected (68%)Expected (95%)
CMS Preliminary -1 = 7 TeV, L = 5.1 fbs -1 = 8 TeV, L = 5.3 fbs ObservedExpected (68%)Expected (95%)
CMS Preliminary -1 = 7 TeV, L = 5.1 fbs -1 = 8 TeV, L = 5.3 fbs FIG. 16: The CMS experiments combined upper limit as afunction of the Higgs boson mass between 100 and 600 GeVSolid black: observed limit/SM; dashed black: CMS expectedlimit/SM in the background-only hypothesis: colored bands: ± , σ distributions around the median expected limit. particle is a boson and plays a role in the mechanism ofelectroweak symmetry breaking.The CMS and ATLAS Collaborations measured sev-eral properties to understand the compatibility of the ob-served boson with the SM Higgs boson and present the re-sults in their papers reporting the observations [199, 201].The experiments fit for the cross section for Higgs bo-son production given the observed data in each decaychannel and globally combining all decay channels. The [GeV] H m110 115 120 125 130 135 140 145 150 ] s Lo c a l S i gn i f i c an c e [ Expected CombinedObserved Combined gg fi Expected H gg fi Observed H llll fi ZZ* fi Expected H llll fi ZZ* fi Observed H n l n l fi WW* fi Expected H n l n l fi WW* fi Observed H bb fi Expected H bb fi Observed H tt fi Expected H tt fi Observed H
ATLAS = 7 TeVs, -1 L dt ~ 4.6-4.8 fb (cid:242) = 8 TeVs, -1 L dt ~ 5.8-5.9 fb (cid:242)
FIG. 17: ATLAS local significance [199] for each search chan-nel and the combination. The observed significance are shownwith solid curves, and the median expected significance as-suming a signal is present at the SM strength are shown withdashed curves. A dash dotted line indicates the 6 σ threshold.The highest local significances of the ZZ and γγ channels are3 . σ and 4 . σ respectively while the combined significance ofall channels is 5 . σ . (GeV) H m
116 118 120 122 124 126 128 130 Lo c a l p - v a l ue -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 σ σ σ σ σ σ σ Combined obs.Expected for SM H γγ→
H ZZ → H WW → H ττ→ H bb → HCMS -1 = 8 TeV, L = 5.3 fbs -1 = 7 TeV, L = 5.1 fbs FIG. 18: CMS local p values [201].The observed p values areshown with solid curves, and the median expected p value forthe combined search assuming a signal is present at the SMstrength is shown with a dashed curve. Horizontal lines indi-cate the 1 σ − σ thresholds. The highest local significancesof the ZZ and γγ channels are 3 . σ and 4 . σ respectivelywhile the combined significance of all channels is 5 . σ . results are presented as a ratio to the expected SM valuesin Figs. 19 and 20 for the ATLAS and CMS experiments,respectively. Of note are the larger than expected cross-section times branching ratios seen in the γγ (ATLASand CMS) and ZZ (ATLAS) decay modes. These modesare dominated by gluon fusion production. The com-bined signal strengths measured by the experiments are1 . ± . . ± .
23 for CMS compatiblewith the SM Higgs boson expectation. Individual signal6 ) m Signal strength ( -1 0 1
Combined 4l fi (*) ZZ fi H gg fi H n l n l fi (*) WW fi H tt fi H bb fi W,Z H -1 Ldt = 4.6 - 4.8 fb (cid:242) = 7 TeV: s -1 Ldt = 5.8 - 5.9 fb (cid:242) = 8 TeV: s -1 Ldt = 4.8 fb (cid:242) = 7 TeV: s -1 Ldt = 5.8 fb (cid:242) = 8 TeV: s -1 Ldt = 4.8 fb (cid:242) = 7 TeV: s -1 Ldt = 5.9 fb (cid:242) = 8 TeV: s -1 Ldt = 4.7 fb (cid:242) = 7 TeV: s -1 Ldt = 5.8 fb (cid:242) = 8 TeV: s -1 Ldt = 4.7 fb (cid:242) = 7 TeV: s -1 Ldt = 4.6-4.7 fb (cid:242) = 7 TeV: s = 126.0 GeV H m – = 1.4 m ATLAS
FIG. 19: ATLAS best-fit signal strength for all SM Higgsboson decays for m H = 125 GeV/ c . SM σ / σ Best fit -1 0 1 2 3 bb → H ττ→ H WW → H ZZ → H γγ→ H CMS -1 = 8 TeV, L = 5.3 fbs -1 = 7 TeV, L = 5.1 fbs = 125.5 GeV H m FIG. 20: CMS best-fit signal strength for all SM Higgs bo-son decays for m H = 125 GeV/ c . Untagged refers to thecross section extracted from topologies sensitive to gluon fu-sion production. The shaded band corresponds to the ± σ uncertainty on the full combination. strengths in the most sensitive modes were discussed inthe sections on individual searches.The fully reconstructed decays of the Higgs boson H → γγ and H → ZZ → ℓ + ℓ − ℓ + ℓ − have excellentmass resolution. The H → W + W − → ℓ + νℓ − ¯ ν decaymode has substantial rate but has poor mass resolutiondue to the two neutrinos in the final state. The ATLASexperiment measures a mass for the observed boson of m H = 126 . ± . stat ) ± . sys ) GeV using all threedecay modes. The individual fits in a two-dimensional [GeV] H m120 125 130 135 140 145 ) m S i gna l s t r eng t h ( Best fit68% CL95% CL gg fi H 4l fi (*) ZZ fi H n l n l fi (*) WW fi H ATLAS -1 Ldt = 4.7-4.8 fb (cid:242) = 7 TeV: s -1 Ldt = 5.8-5.9 fb (cid:242) = 8 TeV: s
FIG. 21: The ATLAS two dimensional fit for the cross sectionand compared to the SM expectation and the Higgs bosonmass for highest significance decay channels.FIG. 22: The CMS two dimensional fit for the cross sectionand compared to the SM expectation and the Higgs bosonmass for highest significance decay channels and the combinedfit using those channels. analysis of signal strength versus mass are shown inFig. 21. The CMS experiment uses the fully recon-structed H → γγ and H → ZZ → ℓ + ℓ − ℓ + ℓ − modes tomeasure a mass of m = 125 . ± . stat ) ± . sys ) GeV.The individual and combined fits are shown in Fig. 22.The results are compatible with limits from previoussearches and the prediction of the SM Higgs boson massfrom constraints derived from electroweak measurements.If the observed boson is involved in the mechanism ofelectroweak symmetry breaking, the measurement of itscoupling to the W and Z bosons is a crucial discrimi-7nant. The production and decay rates measured by theexperiments are compatible with the SM. The ratio ofthe W and Z couplings can be computed by dividingthe production times decay rates for H → W W and H → ZZ since the production of the Higgs boson takesplace via the same mechanisms.The ATLAS experimentmeasures R W Z = 1 . +0 . − . [202] and the CMS experi-ment measures R W Z = 0 . +1 . − . consistent with the SMexpectation where both experiments have normalized themeasurement so that the expected value in the SM is 1.In summary, the LHC experiments extended the LEPexclusion to 122.5 GeV and further excluded a SM Higgsboson with mass between 128 and 600 GeV. The ATLASand CMS experiments both observe a significant excessof events in the region around 125 GeV with evidence forthe production of a new boson in the ZZ and γγ decaymodes, with observed local significances of 4.5 σ and 4.1 σ in the γγ mode and 3.4 σ and 3.1 σ in the ZZ mode.Significant signals (2.8 σ and 1.6 σ ) are also observed inthe H → W W decay mode, while the observed signifi-cance in the fermionic modes ( H → τ τ and H → b ¯ b ) isweak, which is not unexpected given the currently lowexpected significance in these modes. When combiningall their channels, both experiments independently reportthe discovery of a new boson and provide first measure-ments of its fundamental properties, in agreement withthose expected from a SM Higgs boson with a mass closeto 125 GeV. C. Tevatron combined results
As in the LHC experiments, to simplify the combina-tion, the searches are separated into mutually exclusivefinal states. Table IV summarizes for each CDF and D0search the integrated luminosities, the Higgs boson massranges over which the searches are performed, the ratiosof expected and observed limits with respect to SM Higgsboson expectations achieved for m H = 125 GeV, and thereferences to further details for each analysis. Using thecombination procedure outlined in Sec. IX B, limits onSM Higgs boson production σ × B ( H → X ) in p ¯ p colli-sions at √ s = 1 .
96 TeV for 100 ≤ m H ≤
200 GeV areextracted.The combinations of results from each single exper-iment [203, 204], as used in this Tevatron combination,yield the following ratios of 95% C.L. observed (expected)limits to the SM expectation: 2.4 (1.2) for CDF and2.2 (1.6) for D0 at m H = 115 GeV, 2.9 (1.4) for CDFand 2.5 (1.9) for D0 at m H = 125 GeV, and 0.42 (0.69)for CDF and 0.94 (0.76) for D0 at m H = 165 GeV.The ratios of the 95% C.L. expected and observedlimits to the SM cross section are shown in Fig. 23 forthe combined CDF and D0 analyses. The observed (ex-pected) limit values are 1.8 (0.94) at m H = 115 GeV,2.2 (1.1) at m H = 125 GeV, and 0.39 (0.49) at m H =165 GeV. H (GeV/c ) % C L L i m i t / S M Tevatron Run II Preliminary, L ≤ -1 ObservedExpected w/o Higgs ± ± L EP E xc l u s i on Tevatron+ATLAS+CMSExclusion
SM=1 T eva t r on + L EP E xc l u s i on C M S E xc l u s i on A TL A S E xc l u s i on A TL A S E xc l u s i on L EP + A TL A S E xc l u s i on ATLAS+CMSExclusion ATLAS+CMSExclusion
June 2012
FIG. 23: Observed and expected 95% C.L. upper limits onthe ratios to the SM cross section, as functions of the Higgsboson mass for the combined CDF and D0 analyses. Thebands indicate the 68% and 95% probability regions wherethe limits can fluctuate, in the absence of signal. ) Higgs Boson Mass (GeV/c
100 110 120 130 140 150 160 170 180 190 200 B a ck g r ound p - v a l ue -5 -4 -3 -2 -1 s s s s Tevatron RunII Preliminary
June 2012 -1 £ L Observed b b – Expected 2 s.d. – Expected
Tevatron RunII Preliminary -1 £ L FIG. 24: The background p values 1-C . L . b as a function ofthe Higgs boson mass (in steps of 5 GeV), for the combinationof the CDF and D0 analyses. The green and yellow bandscorrespond, respectively, to the regions enclosing 1 σ and 2 σ fluctuations around the median prediction in the signal-plus-background hypothesis at each value of m H . Fig. 24 shows the p value 1-CL b as a function of m H ,i.e., the probability that an upward fluctuation of thebackground can give an outcome as signal-like as thedata or more. In the absence of signal, the p value isexpected to be uniformly distributed between 0 and 1. Asmall p value indicates that the data are unlikely to beexplained by the background-only hypothesis. The small-est observed p value corresponds to a Higgs boson massof 120 GeV and has a local significance of 3.0 σ . Thefluctuations seen in the observed p value as a functionof the tested m H result from excesses seen in different8 TABLE IV: The integrated luminosity, explored mass range, 95% C.L. expected and observed limits on Higgs boson productioncross section relative to the SM expectation ( R ) assuming m H = 125 GeV, and references for the different CDF and D0 analysesgrouped by the final states ( ℓ = e or µ ) considered.CDF Channels Luminosity (fb − ) m H range exp. R obs. R Reference2 TeV (GeV) 125 GeV 125 GeV
W H → ℓνb ¯ b −
150 2.8 4.9 [113] ZH → ℓ + ℓ − b ¯ b −
150 3.6 7.2 [117] ZH → ν ¯ νb ¯ b −
150 3.6 6.8 [121] H → W + W − & W H → W W + W − & ZH → ZW + W − −
200 3.1 3.0 [145] H → γγ −
150 9.9 17.0 [129] H → ZZ (four leptons, limits are given at 130 GeV) 9.7 120 −
200 18.3 20.5 [151] H → W + W − ( eτ h )+( µτ h ) & W H → W W + W − (1 τ h ) 9.7 130 −
200 . . [150] H + X → τ + τ − (1 jet)+(2 jet) 8.3 100 −
150 14.8 11.7 [125]
W H → ℓντ + τ − / ZH → ℓ + ℓ − τ + τ − −
150 23.3 26.5 [126]
W H + ZH → jjb ¯ b (SS,SJ) 9.5 100 −
150 9.0 11.0 [107] t ¯ tH → W W b ¯ bb ¯ b (no lepton) - (lepton) 5.7-9.5 100 −
150 12.4 17.6 [131, 132]D0 Channels Luminosity (fb − ) m H range exp. R obs. R Reference2 TeV (GeV) 125 GeV 125 GeV
W H → ℓνb ¯ b −
150 4.7 5.2 [114] ZH → ℓ + ℓ − b ¯ b −
150 5.1 7.1 [118] ZH → ν ¯ νb ¯ b −
150 3.9 4.3 [122] H → W + W − → ℓ ± νℓ ∓ ν −
200 3.6 4.6 [146] H → γγ −
150 8.2 12.9 [130] H → W + W − → µντ h ν −
200 12.8 15.7 [127] H → W + W − → ℓ ¯ νjj −
200 . . [149] H + X → ℓ ± τ ∓ h jj −
200 40.0 44.0 [127]
V H → τ τ µ + X −
200 17.6 13.1 [128]
V H → e ± µ ± + X −
200 11.6 7.8 [147]
V H → ℓℓℓ + X −
200 11.1 19.3 [148] search channels, as well as from point-to-point fluctua-tions originating from the separate discriminants used ateach m H , as next discussed in more detail. The widthof the dip from 115 to 135 GeV is consistent with thecombined resolution of the H → b ¯ b and H → W + W − channels. The effective resolution of this search comesfrom two independent sources.: the reconstructed candi-date masses, which directly constrain m H , and the ex-pected cross sections times the relevant branching ratiosfor the H → b ¯ b and H → W + W − channels, which arefunctions of m H in the SM. The observed excess in the H → b ¯ b channels coupled with a less signal-like outcomein the H → W + W − channels determines the shape ofthe observed p value as a function of m H .The strongest sensitivity at low mass comes from the H → b ¯ b channels. The largest local significance in thecombination of H → b ¯ b channels is 3.3 σ at a mass of135 GeV, while it is 2.8 σ at 125 GeV [108].In Fig. 25, the signal strength is allowed to vary as afunction of m H in the fit of the signal-plus-backgroundhypothesis to the observed data over the full mass range.As shown, the resulting best-fit signal strength normal-ized to the SM prediction is within 1 σ of the SM ex-pectation for a Higgs boson signal in the range 110 140 GeV. The largest signal fit in this range, nor- H (GeV/c ) B e s t f i t s / S M Tevatron Run II Preliminary, L ≤ -1 June 2012 SM=1 FIG. 25: The best-fit signal cross section of all CDF and D0search channels combined shown as a ratio to the SM crosssection as a function of the tested Higgs boson mass. Thehorizontal line at 1 represents the signal strength expectedfor a SM Higgs boson hypothesis. The blue band shows the1 σ uncertainty on the signal fit. malized to the SM prediction, is obtained at 130 GeV,rather than for the smallest p -value mass of 120 GeV,9since the similar excesses for these two mass hypothe-ses translate into a higher signal strength at 130 GeV.The excess in signal-strength around 200 GeV occurs ina region of low expected sensitivity ( ∼ σ ) and with anunphysically narrow mass range; thus it cannot be at-tributed to a SM Higgs boson signal at high mass.At the Tevatron the look-elsewhere effect (LEE) is es-timated in a simplified and conservative manner. In themass range 115–150 GeV, where the low-mass H → b ¯ b searches dominate, the reconstructed mass resolution isapproximately 15%. A LEE factor of ≃ H → γγ searcheshave a much better mass resolution, of the order of 3%,but their contribution to the final LLR is small dueto the much smaller signal-to-background ratio in thosesearches. The H → τ + τ − searches have both worse re-constructed mass resolution and lower signal to back-ground ratio than the H → b ¯ b searches, and thereforesimilarly do not play a significant role in the estimationof the LEE. The H → W W channel has the poorest massresolution and therefore contributes weakly to the LEE.For the combined search of all Tevatron channels, witha conservative LEE of ≃ σ H (GeV/c ) ( s W H + s Z H ) x B r( H → bb – ) ( f b ) Tevatron Run II, L int ≤ -1 Measured ± ± H =125 GeV/c Assuming best fit rate at m H =125 GeV/c Expected for m H =125 GeV/c Assuming SM rate FIG. 26: The best-fit cross section times branching ratio σ WH + σ ZH × B ( H → b ¯ b ) as a function of m H measured atthe Tevatron. Applying the low-mass LEE to the most significantlocal p value obtained from the CDF+D0 H → b ¯ b com-bination, a global significance of approximately 3.1 σ isobtained, resulting in evidence for the production of aresonance in the b -flavored dijet mass distribution, pro-duced in association with a massive vector boson. Giventhe mass resolution in this final state, this resonance isconsistent with the new boson observed by the LHC ex-periments, and provides the first evidence for fermionicdecays of this boson.The measured cross section times branching ratio σ W H + σ ZH × B ( H → b ¯ b ) is shown in Fig. 26 as a func-tion of m H . The resulting value is 0 . +0 . − . pb for m H = 125 GeV, consistent with the corresponding SM pre-diction of 0.12 ± m H = 125 GeV. SM s / s Best Fit b b fi H gg fi H - W + W fi H = 125 GeV/c H m Combined (68%)Single Channel Tevatron Run II Preliminary -1 £ L June 2012 FIG. 27: Best-fit signal strength for the three most sensitiveboson decay modes at the Tevatron, for m H = 125 GeV/ c .The shaded band corresponds to the ± σ uncertainty on thefull combination. In summary, at the Tevatron, when combining allsearch channels, there is significant excess of data eventswith respect to the background estimation in the massrange 115 < m H < 135 GeV. The p value for a back-ground fluctuation to produce such an excess correspondsto a local significance of 3.0 σ at 120 GeV. The largestexcess is observed in the H → b ¯ b channels, with a localsignificance of 3.3 σ , which results in a global significanceof ≈ . σ when accounting for look-elsewhere effects.The CDF and D0 Collaborations thus report evidencefor the production of a resonance in the b -flavored dijetmass distribution produced in association with a massivevector boson, consistent with the new boson observed bythe LHC Collaborations. The measured cross section forthis process is consistent with the cross section expectedfor a SM Higgs boson of 125 GeV produced in associationwith a W or a Z boson. D. Conclusion and prospects The LHC experiments have discovered a new bosonwith mass around 125 GeV and have evidence for thisparticle in several decay modes. The Tevatron exper-iments report evidence for a particle, produced in as-sociation with W or Z bosons and which decays to b ¯ b ,with a mass compatible to that reported by the LHC ex-periments. The production and decay modes that have0been observed indicate that this boson plays a role in themechanism of electroweak symmetry breaking and alsoin the mass generation for the quarks. The propertiesof this boson are compatible with those expected for aSM Higgs boson but more study is required to fully ex-plore the nature of this discovery. The discovery of a newboson with properties indicating that it plays a role inelectroweak symmetry breaking is a major breakthroughin fundamental physics.The LHC experiments expect to integrate up to 30 fb − of data at 8 TeV center of mass energy by late 2012.These data should be sufficient to make first measure-ments of all accessible parameters of the boson assuming SM-like behavior. After the 2012 run the LHC is ex-pected to undergo a long shutdown to upgrade the energyand luminosity capabilities of the accelerator, to near thedesign parameters of 14 TeV and 100 fb − per year. Datataken after the upgrades should allow for precision mea-surements of the boson’s properties and exploration ofnon SM physics associated with the boson. After completion of this review, additional data takenby the ATLAS and CMS collaborations confirmed that thenew discovered boson is a Higgs boson, while the currentprecision of its measured properties does not yet allowone to definitely identify it as the standard model Higgsboson. [1] P.W. Higgs, Phys. Rev. Lett. 12, 132 (1964), idem , Phys.Rev. 145, 1156 (1966)[2] F. Englert and R. Brout, Phys. Rev. Lett. 13, 321 (1964)[3] G.S. Guralnik, C.R. Hagen, and T.W. Kibble, Phys.Rev. Lett. 13, 585 (1964)[4] S.L. Glashow, Nucl. Phys. 22, 579 (1961)[5] S. Weinberg, Phys. Rev. Lett. 19, 1264 (1967)[6] A. Salam, Elementary Particle Theory , ed: N.Svartholm, Almquist, and Wiksells (1968)[7] S.L. Glashow, J. Iliopoulos, and L. Maiani, Phys. Rev.D 2, 1285 (1970)[8] J.F. Gunion, H.E. Haber, G.L. Kane, and S. Dawson, The Higgs Hunter’s Guide (Addison-Wesley) (1990)[9] A. Djouadi, M. Spira, P.M. Zerwas, Phys. Lett. B264,440 (1991)[10] S. Dawson, Nucl. Phys. B359, 283 (1991)[11] R.V. Harlander and W.B. Kilgore, Phys. Rev. Lett. 88,201801 (2002) arXiv:hep-ph/0201206[12] C. Anastasiou and K. Melnikov, Nucl. Phys. B646, 220(2002) arXiv:hep-ph/0207004[13] V. Ravindran, J. Smith, W.L. van Neerven, Nucl. Phys.B665, 325 (2003) arXiv:hep-ph/0302135[14] S. Actis, G. Passarino, C. Sturm, S. Uccirati, Phys. Lett.B670, 12 (2008) arXiv:0809.1301 [hep-ph][15] U. Aglietti, R. Bonciani, G. Degrassi, A. Vicini, Phys.Lett. B595, 432 (2004) arXiv:hep-ph/0404071[16] G. Degrassi and F. Maltoni, Phys. Lett. B600, 255(2004) arXiv:hep-ph/0407249[17] C. Anastasiou, R. Boughezal and F. Petriello, JHEP0904, 003 arXiv:0811.3458 (2009)[18] M. Kramer, E. Laenen and M. Spira, Nucl. Phys. B511,523 (1998) arXiv:hep-ph/9611272[19] K.G. Chetyrkinm B.A. Kniehl, M. Steinhauser, Nucl.Phys. B510, 61 (1998) arXiv:hep-ph/9708255[20] D. de Florian, M. Grazzini and Z. Kunszt, Phys. Rev.Lett. 82, 5209 (1999)[21] V. Ravindran, J. Smith and W. L. Van Neerven, Nucl.Phys. B , 247 (2002) arXiv:0201114 [hep-ph][22] S. Catani, D. de Florian, M. Grazzini, P. Nason, JHEP0307, 028 (2003) arXiv:hep-ph/0306211[23] S. Moch and A. Vogt, Phys. Lett. B631, 48 (2005)[24] E. Laenen and L. Magnea, Phys. Lett. B632, 270 (2006)arXiv:hep-ph/0508265[25] A. Idilbi, X. d. Ji, J. P. Ma, F Yuan, Phys. Rev. D 73,077501 (2006) arXiv:hep-ph/0509294 [26] V. Ravindran, Nucl. Phys. B746, 58 (2006)arXiv:hep-ph/0512249[27] D. de Florian and M. Grazzini, Phys. Lett. B674, 291(2009) arXiv:0901.2427 [hep-ph][28] J. Baglio and A. Djouadi, JHEP 1103 055 (2011)arXiv:1012.0530 [hep-ph][29] C. Anastasiou, S. Buehler, F. Herzog, and A. Lazopou-los, JHEP 1204 004 (2012) arXiv:1202.3638 [hep-ph][30] D. de Florian, G. Ferrera, M. Grazzini, and D. Tom-masini, JHEP 1111 064 (2011) arXiv:1109.2109 [hep-ph][31] E. Bagnaschi, G. Degrassi, P. Slavich and A. Vicini,JHEP 1202 88 (2012) arXiv:1111.2854 [hep-ph][32] C. R. Schmidt, Phys. Lett. B413,391(1997)[33] C.J. Glosser and C.R. Schmidt, JHEP , 016(2002)[34] J. M. Campbell, R. K. Ellis, and G. Zanderighi, JHEP , 028 (2006)[35] J. M. Campbell, R. K. Ellis, and C. Williams, Phys.Rev. D81,074023(2010)[36] K.A.Assamagan et al. , (2004) arXiv:hep-ph/0406152[37] O. Brein, A. Djouadi, R. Harlander, Phys. Lett. B579,149 (2004) arXiv:hep-ph/0307206[38] M.L. Ciccolini, S. Dittmaier, M. Kr¨amer, Phys. Rev.D68, 073003 (2003) arXiv:hep-ph/0306234[39] J. Baglio and A. Djouadi, J. High Energy Phys. 1010,064 arXiv:1009.1363 (2010)[40] G. Ferrera, M. Grazzini and F. Tramontano, Phys. Rev.Lett. 107, 152003 arXiv:1107.1164 (2011)[41] A. Denner, S. Dittmaier, S. Kallweit and A. Muck,arXiv:1112.5142 [hep-ph] (2011)[42] O. Brein, R. Harlander, M. Wiesemann and T. Zirke(2011) arXiv:1111.0761 [hep-ph][43] T. Figy, S. Palmer, G. Weiglein, JHEP 1202, 105 (2012)arXiv:1012.4789 [hep-ph][44] P. Bolzoni, F. Maltoni, S.-O. Moch, M. Zaro, Phys. Rev.Lett. 105, 011801 (2010) arXiv:1003.4451 [hep-ph][45] M. Ciccolini, A. Denner, S. Dittmaier, Phys. Rev. Lett.99, 161803 (2007) arXiv:0707.0381 [hep-ph][46] Ciccolini, A. Denner, S. Dittmaier, Phys. Rev. D 77,103002 arXiv:0710.4749 [hep-ph] (2008)[47] T. Han, G. Valencia, S. Willenbrock, Phys. Rev. Lett.69, 3274 arXiv:hep-ph/9206246 (1992)[48] E.L. Berger, J. Campbell, Phys. Rev. D 70, 073011arXiv:hep-ph/0403194 (2004)[49] T. Figy, C. Oleari, D. Zeppenfeld, Phys. Rev. D 68,073005 arXiv:hep-ph/0306109 (2003) [50] S. Dittmaier et. al. , [LHC Higgs Cross Section WorkingGroup], (2012) arXiv:1201.3084 [hep-ph][51] W. Beenakker, S. Dittmaier, M. Kr¨amer, B. Pl¨umper,M. Spira and P.M. Zerwas, Phys. Rev. Lett. 87, 201805(2001) arXiv:hep-ph/0107081[52] L. Reina and S. Dawson, Phys. Rev. Lett. 87, 201804(2001)[53] S. Dawson, L.H. Orr, L. Reina, D. Wackeroth, Phys.Rev. D 67, 071503 (2003) arXiv:hep-ph/0107101[54] W. Beenakker, S. Dittmaier, M. Kr¨amer, B. Pl¨umper,M. Spira and P.M. Zerwas, Nucl. Phys. B653, 151 (2003)arXiv:hep-ph/0211352[55] S. Dittmaier et. al. , [LHC Higgs Cross Sections WorkingGroup], (2011) arXiv:1101.0593 [hep-ph][56] For a compilation of theoretical results for SM andMSSM Higgs cross sections at the LHC see: https://twiki.cern.ch/twiki/bin/view/LHCPhysics/CrossSections .For the SM Higgs production via gluon fusion us-ing Renormalization Group improved predictions see: http://rghiggs.hepforge.org/ (2012).[57] LEP Electroweak Working Group, status of March(2012), http://lepewwg.web.cern.ch/LEPEWWG/ , TheALEPH, CDF, DØ, DELPHI, L3, OPAL, SLD Collabo-rations, the LEP Electroweak Working Group, the Teva-tron Electroweak Working Group, and the SLD Elec-troweak and Heavy Flavor groups, LEPEWWG/2009-01 (2009) arXiv:0911.2604 [hep-ex] ,[58] J. Erler and P. Langacker, Phys.Lett. B592 1 (2004)arXiv:hep-ph/0407097[59] CDF and D0 Collaborations, and the Tevatron Elec-troweak Working Group, (2011) arXiv:1107.5255[60] Tevatron Electroweak Working Group, CDF and D0Collaborations, (2012) arXiv:1204.0042 [hep-ex][61] B.L. Ioffe and V.A. Khoze, Sov. J. Part. Nucl. Phys. 9,50 (1978)[62] J. Ellis et. al. , Nucl. Phys. B106, 292 (1976)[63] D.R.T. Jones and S.T. Petcov, Phys. Lett. 84B, 440(1979)[64] R.N. Cahn and S. Dawson, Phys. Lett. 136B, 196 (1984), ibid. , 138B, 464 (1984)[65] W. Kilian, M. Krmer, P. Zerwas, Phys. Lett. B373, 135(1996) arXiv:hep-ph/9512355[66] ALEPH, DELPHI, L3, and OPAL Collaborations, TheLEP Working Group for Higgs Boson Searches, Phys.Lett. B565, 61 (2003) arXiv:hep-ex/0306033[67] CDF and D0 Collaborations, Phys. Rev. Lett. 104,061802 (2010) arXiv:1001.4162 [hep-ex][68] CDF and D0 Collaborations (2010 b) arXiv:1007.4587[hep-ex][69] CDF and D0 Collaborations (2011) arXiv:1107.5518[hep-ex][70] GFitter collaboration, M. Baak, et al. , Eur.Phys.J. C722003 (2012) arXiv:1107.0975 [hep-ph][71] M. L. Mangano, M. Moretti, F. Piccinini and M. Trec-cani, JHEP , 013 (2007) arXiv:hep-ph/0611129[72] J. Alwall et al. , JHEP , 128 (2011) arXiv:hep-ph/1106.0522[73] P. Nason, JHEP 040 (2004) hep-ph/0409146[74] S. Frixione, P. Nason and C. Oleari, JHEP , 070(2007) arXiv:0709.2092[75] S. Alioli, P. Nason, C. Oleari and E. Re, JHEP ,043 (2010) arXiv:1002.2581[76] T. Sj¨ostrand, S. Mrenna, and P. Skand, JHEP , 026 (2006) arXiv:hep-ph/0603175[77] CDF Collaboration, Phys. Rev. Lett. 103, 092002 (2009)arXiv:0903.0885 [hep-ex][78] D0 Collaboration, Phys. Rev. Lett.103, 092001 (2009)arXiv:0903.0850 [hep-ex][79] E. Boos, V. Bunichev, M. Dubinin, L. Dudko, V. Ilyin,A. Kryukov, V. Edneral, V. Savrin, A. Semenov, andA. Sherstnev, Nucl. Instrum. Methods Phys. Res., Sect.A , 250 (2004)[80] E. E. Boos, V. E. Bunichev, L. V. Dudko, V. I. Savrin,and A. V. Sherstnev, Phys. Atom. Nucl. , 1317(2006).[81] S. Frixione and B.R. Webber, J. High Energy Phys. 029 (2002).[82] T. Binoth et al. , JHEP , 046 (2006)arXiv:hep-ph/0611170[83] T. Binoth, N. Kauer, P. Mertsch (2008) arXiv:hep-ph/0807.0024[84] CDF Collaboration, J. Phys. G G 34, 2457 (2007)arXiv:hep-ex/0508029[85] D0 Collaboration, Nucl. Instrum. Methods, Phys. Res.A 565, 463 (2006) arXiv:physics/0507191[86] For a recent review, see P. C. Bhat, Ann. Rev. Nucl.Part. Sci. , 281 (2011). The specific details of theMVA for each analysis are described in the respectivereferences.[87] D0 Collaboration, Nucl. Instrum. Methods, Phys. Res.A 620, 490 (2010)[88] CDF Collaboration, Nucl.Instrum.Meth. A697 64(2013) arXiv:1205.1812 [hep-ex][89] CDF Collaboration, Phys. Rev. D 72, 052003 (2005)arXiv:hep-ex/0504053[90] ATLAS Collaboration, JINST S08003 (2008)http://inspirehep.net/record/796888/.[91] CMS Collaboration, JINST S08004 (2008)http://inspirehep.net/record/796887.[92] S. Baffioni et al. , Eur. Phys. J. C 49, 1099 (2007)http://inspirehep.net/record/713909[93] CMS Collaboration, CMS-PAS-EGM-10-004 (2010)http://cdsweb.cern.ch/record/1299116.[94] ATLAS Collaboration, JHEP 1012, 060 (2010)arXiv:1010.2130 [hep-ex][95] CMS Collaboration, CMS-PAS-PFT-10-003 (2010 b)http://cds.cern.ch/record/1279347[96] CMS Collaboration, CMS-PAS-MUO-10-002 (2010 c)http://cdsweb.cern.ch/record/1279140.[97] CMS Collaboration, JINST 7 P01001 (2012)arXiv:1109.6034 [physics][98] ATLAS Collaboration, ATLAS-CONF-2011-152 (2011)http://cdsweb.cern.ch/record/1398195.[99] ATLAS Collaboration, ATL-PHYS-PUB-2011-007(2011 b)http://cdsweb.cern.ch/record/1345329.[100] M. Cacciari, G. P. Salam and G. Soyez, JHEP 0804, 063(2008) arXiv:0802.1189 [hep-ph][101] CMS Collaboration, CMS-PAS-BTV-10-001 (2010 d)http://cdsweb.cern.ch/record/1279144.[102] ATLAS Collaboration, ATLAS-CONF-2011-102 (2011c)http://cdsweb.cern.ch/record/1369219.[103] CMS Collaboration, CMS-PAS-PFT-09-001 (2009)http://cdsweb.cern.ch/record/1194487.[104] ATLAS Collaboration, Eur. Phys. J. C 72, 1844 (2012)arXiv:1108.5602 [hep-ex] [105] CMS Collaboration, CMS-PAS-PFT-09-001 (2009 b)http://cdsweb.cern.ch/record/1194487.[106] CMS Collaboration, CMS-PAS-PFT-10-002 (2010 e)http://cds.cern.ch/record/1279341[107] CDF Collaboration, JHEP 1302 004 (2013 b)arXiv:1208.6445 [hep-ex][108] CDF and D0 Collaborations, Phys. Rev. Lett. 109,071804 (2012) arXiv:1203.3782 [hep-ex][109] CDF Collaboration, Phys.Rev.Lett. 109, 111804 (2012)arXiv:1207.1703 [hep-ex][110] D0 Collaboration, Phys.Rev.Lett. 109, 121802 (2012)arXiv:1207.6631 [hep-ex][111] CDF Collaboration, Phys.Rev.Lett. 103, 101802 (2009b) arXiv:0906.5613 [hep-ex][112] D0 Collaboration, Phys. Rev. Lett. 102, 051803 (2009b) arXiv:0808.1970 [hep-ex][113] CDF Collaboration, Phys.Rev.Lett. 109, 111804 (2012b) arXiv:1207.1703 [hep-ex][114] D0 Collaboration, Phys. Rev. Lett. 109, 121804 (2012b) arXiv:1208.0653 [hep-ex][115] CDF Collaboration, Phys.Lett. B715, 98 (2012 c)arXiv:1203.5815 [hep-ex][116] D0 Collaboration, Phys. Rev. Lett. 105, 251801 (2010b) arXiv:1008.3564 [hep-ex][117] CDF Collaboration, Phys.Rev.Lett. 109, 111803 (2012d) arXiv:1207.1704 [hep-ex][118] D0 Collaboration, Phys. Rev. Lett. 109, 121803 (2012c) arXiv:1207.5819 [hep-ex][119] CDF Collaboration, Phys.Rev.Lett. 104, 131801 (2010)arXiv:0911.3935 [hep-ex][120] D0 Collaboration, Phys. Rev. Lett. 104, 071801 (2010c) arXiv:0912.5285 [hep-ex][121] CDF Collaboration, Phys.Rev.Lett. 109, 111805 (2012e) arXiv:1207.1711 [hep-ex][122] D0 Collaboration, Phys. Lett. B 716, 285 (2012 d)arXiv:1207.5689 [hep-ex][123] CDF Collaboration, Phys. Rev. Lett. 109, 111802 (2012f) arXiv:1207.1707 [hep-ex][124] D0 Collaboration, Phys. Rev. Lett. 109, 121802 (2012e) arXiv:1207.6631 [hep-ex][125] CDF Collaboration, “Search for the standard modelHiggs boson in τ + τ − plus jets final state with 8.3 fb − of CDF data”, CDF Conference Note 10625 (2011)[126] CDF Collaboration, “Search for the standard modelHiggs in the ℓντ τ and ℓℓτ τ channels”, CDF ConferenceNote 10500 (2011 b)[127] D0 Collaboration, “Search for the standard model Higgsboson in tau pair final states”, D0 Conference note 6305(2012 f).[128] D0 Collaboration, “Search for a standard model Higgsboson in the τ τ µ final state with 7.0 fb − at √ s =1 . 96 TeV”, D0 Conference Note 6286 (2012 g).[129] CDF Collaboration, Phys.Lett. B717, 173 (2012 g)arXiv:1207.6386 [hep-ex][130] D0 Collaboration, “Search for the Standard ModelHiggs boson in γγ + X final states at D0 with 9.7 fb − of data”, D0 Conference Note 6295 (2012 h).[131] CDF Collaboration, Phys.Rev.Lett. 109 181802 (2012h) arXiv:1208.2662 [hep-ex][132] CDF Collaboration, “Search for standard model Higgsboson production in association with t ¯ t using no leptonfinal state”, CDF Conference Note 10582 (2011 c)[133] CDF Collaboration, Phys. Rev. Lett. 78, 4536http://inspirehep.net/record/423258/ (1997) [134] D0 Collaboration, Phys. Rev. Lett. 94, 151801 (2005)arXiv:hep-ex/0410066, [Erratum-ibid. 100, 139901(2008)][135] CDF Collaboration, Phys. Rev. Lett. 98, 161801 (2007b) arXiv:hep-ex/0702027][136] D0 Collaboration, Phys. Rev. Lett. 101, 171803 (2008)arXiv:0808.0703 [hep-ex][137] D0 Collaboration, Phys. Rev. Lett. 103, 191801 (2009c) arXiv:0904.0673 [hep-ex][138] CDF Collaboration, Phys. Rev. Lett. 104, 201801 (2010b) arXiv:0912.4500 [hep-ex][139] CDF Collaboration, Phys. Rev. D86, 031104 (2012 i)arXiv:1202.6629 [hep-ex][140] D0 Collaboration, Phys.Rev. D85, 112005 (2012 i)arXiv:1201.5652 [hep-ex][141] CDF Collaboration, Phys. Rev. Lett. 108, 101801 (2012j) arXiv:1112.2978 [hep-ex][142] CDF Collaboration, Phys. Rev. Lett. 104, 101801 (2010c) arXiv:0911.4449 [hep-ex][143] D0 Collaboration, Phys. Rev. Lett. 102, 161801 (2009d) arXiv:0810.3873 [hep-ex][144] CDF Collaboration, Phys. Rev. Lett. 103, 091803 (2009c) arXiv:0905.4714 [hep-ex][145] CDF Collaboration, “Search for H → W W ∗ productionusing 9.7 fb − ”, CDF Conference Note 10785 (2012 k)[146] D0 Collaboration, “ Search for Higgs boson productionin dilepton plus missing energy final states with 8.6 - 9.7fb − of p ¯ p collisions at √ s = 1 . 96 TeV”, D0 ConferenceNote 6302 (2012 j).[147] D0 Collaboration, “Search for associated Higgs bosonproduction with V H → e ± ν e µ ± ν µ + X like chargedelectron muon pairs using 9.7 fb − of p ¯ p collisions at √ s = 1 . 96 TeV”, D0 Conference Note 6301 (2012 k).[148] D0 Collaboration, “Search for standard model Higgs bo-son with tri-leptons and missing transverse energy with9.7 fb − of p ¯ p collisions at √ s = 1 . 96 TeV”, D0 Confer-ence Note 6276 (2012 l).[149] D0 Collaboration, Phys. Rev. Lett. 106, 171802 (2011)arXiv:1001.6079v2 [hep-ph][150] CDF Collaboration, “Search for H → W W ∗ with lep-tons and hadronic taus in the final state using 9.7 fb − c)http://cdsweb.cern.ch/record/1430734.[161] CMS Collaboration, CMS-PAS-SMP-12-005 (2012 c)http://cdsweb.cern.ch/record/1440234.[162] CMS Collaboration, CMS-PAS-SMP-12-013 (2012 d)http://cdsweb.cern.ch/record/1460099.[163] ATLAS Collaboration, Phys. Rev. Lett. 108, 041804(2012 d) arXiv:1110.5016 [hep-ex][164] CMS Collaboration, JHEP 01 063 (2013)arXiv:1211.4890 [hep-ex][165] ATLAS Collaboration, ATLAS-CONF-2012-090 (2012e)http://cdsweb.cern.ch/record/1460409[166] CMS Collaboration, Submitted to Phys. Lett. B (2013b) arXiv:1301.4698 [hep-ex][167] ATLAS Collaboration, Phys. Rev. Lett. 108, 111803(2012 f) arXiv:1202.1414 [hep-ex][168] ATLAS Collaboration, ATLAS-CONF-2012-091 (2012g)http://cds.cern.ch/record/1460410[169] CMS Collaboration, Phys. Lett. B 710, 403 (2012 e)arXiv:1202.1487 [hep-ex][170] CMS Collaboration, CMS-PAS-HIG-12-015 (2012 f)http://cdsweb.cern.ch/record/1460419.[171] ATLAS Collaboration, Phys. Lett. B 710, 383 (2012 h)arXiv:1202.1415 [hep-ex][172] ATLAS Collaboration, ATLAS-CONF-2012-092 (2012i)http://cdsweb.cern.ch/record/1460411.[173] CMS Collaboration, Phys.Rev.Lett. 108, 111804 (2012g) arXiv:1202.1997 [hep-ex][174] CMS Collaboration, CMS-PAS-HIG-12-016 (2012 h)http://cdsweb.cern.ch/record/1460664.[175] CMS Collaboration, JHEP 1203, 081 (2012 i)arXiv:1202.3617 [hep-ex][176] CMS Collaboration, CMS-PAS-SMP-12-009 (2012 j)http://cdsweb.cern.ch/record/1431862.[177] ATLAS Collaboration, Phys. Lett. B, 716, 62 (2012 j)arXiv:1206.0756 [hep-ex][178] ATLAS Collaboration, ATLAS-CONF-2012-098 (2012k)http://cdsweb.cern.ch/record/1462530.[179] CMS Collaboration, CMS-PAS-HIG-12-017 (2012 k)http://cdsweb.cern.ch/record/1460424.[180] CMS Collaboration, Phys. Lett. B 710, 91 (2012 l)arXiv:1202.1489 [hep-ex][181] ATLAS Collaboration, Phys.Lett. B718, 391 (2012 l)arXiv:1206.6074 [hep-ex].[182] CMS Collaboration, CMS-HIG-12-003 (2012 m)http://cds.cern.ch/record/1449158[183] CMS Collaboration, CMS-HIG-12-021 (2012 n) http://cdsweb.cern.ch/record/1460660[184] ATLAS Collaboration, Phys.Lett. B717, 70 (2012 m)arXiv:1206.2443 [hep-ex].[185] CMS Collaboration, JHEP 1204, 036 (2012 o)arXiv:1202.1416 [hep-ex][186] ATLAS Collaboration, Phys.Lett. B717, 29 (2012 n)arXiv:1205.6744 [hep-ex][187] CMS Collaboration, JHEP 1203, 040 (2012 p)arXiv:1202.3478 [hep-ex][188] CMS Collaboration, CMS-PAS-HIG-12-023 (2012 q)http://cdsweb.cern.ch/record/1460693.[189] ATLAS Collaboration, JHEP 1209, 070 (2012 o)arXiv:1206.5971 [hep-ex][190] CMS Collaboration, Phys. Lett. B 713, 68 (2012 r)arXiv:1202.4083 [hep-ex][191] CMS Collaboration, CMS-PAS-HIG-12-018 (2012 s)http://cdsweb.cern.ch/record/1460413.[192] ATLAS Collaboration, Phys.Lett. B718, 369 (2012 p)arXiv:1207.0210 [hep-ex][193] CMS Collaboration, Phys. Lett. B 710, 284 (2012 t)arXiv:1202.4195 [hep-ex][194] CMS Collaboration, CMS-PAS-HIG-12-019 (2012 u)http://cdsweb.cern.ch/record/1460692.[195] ATLAS Collaboration, ATLAS-CONF-2012-135 (2012q)http://cdsweb.cern.ch/record/1478423.[196] CMS Collaboration, CMS-PAS-HIG-12-020 (2012 v)http://cdsweb.cern.ch/record/1460438.[197] ATLAS Collaboration, ATLAS-CONF-2012-019 (2012r)http://cds.cern.ch/record/1430033[198] ATLAS Collaboration, Phys. Lett. B 710, 49 (2012 s)arXiv:1202.1408 [hep-ex][199] ATLAS Collaboration, Phys. Lett. B 716, 1 (2012 t)arXiv:1207.7214 [hep-ex][200] CMS Collaboration, Phys. Lett. B 710, 26 (2012 w)arXiv:1202.1488 [hep-ex][201] CMS Collaboration, Phys. Lett. B 716, 30 (2012 x)arXiv:1207.7235 [hep-ex][202] ATLAS Collaboration, ATLAS-CONF-2012-127 (2012u)http://cds.cern.ch/record/1476765[203] CDF Collaboration, “Combination of CDF standardmodel Higgs boson searches with up to to 9.7 fb − ofdata”, CDF Conference Note 10804 (2012 o)[204] D0 Collaboration, “Combined Upper Limits on Stan-dard Model Higgs Boson Production from the D0 Ex-periment in up to 9.7 fb −1