IPPP/11/73 DCPT/11/146 LPN11-61 Evasive Higgs Maneuvers at the LHC Christoph Englert,1,2,∗ Joerg Jaeckel,2,† Emanuele Re,2,‡ and Michael Spannowsky2,§ 1Institut fu¨r Theoretische Physik, Universita¨t Heidelberg, 69120 Heidelberg, Germany 2Institute for Particle Physics Phenomenology, Department of Physics, Durham University, DH1 3LE, United Kingdom Non-standard decays of the Higgs boson produced at the Large Hadron Collider can lead to signatures which can easily be missed due to non-adapted trigger or search strategies. Keeping electroweak symmetry breaking Standard Model-like we classify the phenomenology of an evasive Higgsbosonintothreecategoriesanddiscusshowtheycanbedescribedinaneffectivefieldtheory. Wecomment on how one can improvethe search strategies to also detect such an evasive Higgs. 2 I. INTRODUCTION how “hide” from the Standard searches. 1 0 The relevant features for a Higgs search are the pro- 2 duction rate and the decay signatures in the detector. One of the main goals of the Large Hadron Collider Therefore one option to hide the Higgs is to try and sig- n (LHC) is the discovery of the Higgs boson responsible a for electroweak symmetry breaking [1]. With the LHC nificantly reduce the production of the Higgs. Alterna- J having provided about 5 fb−1 of data, both Atlas and tively, the Higgs could dominantly decay into particles 8 Cms have presented strong constraints on a Standard which are difficult to detect or difficult to disentangle 1 from the background. Model (SM) Higgs boson. If the Higgs is SM-like it is disfavored at the 95% confidence level in the mass Constraints on our ability to reduce the production ] h range131GeV(127GeV)to453GeV(600GeV)byAt- cross section arise from the fact that Higgs production p las(Cms)[2,3],whiletheLEP2bound[4]of114.4GeV rates in the SM are bound to the physical Higgs being - is pushed to 115.5 GeV (115 GeV). With the rapidly the unitarizing degree of freedom in longitudinal gauge p e growing data sample we can expect to have full cover- boson scattering VLVL VLVL (V = W±,Z) and in h ageofthe SMHiggsmass atthe 95%confidence levelall massive qq¯ VLVL am→plitudes∗. In the SM this fixes → [ the way down to 115 GeV at the LHC for an integrated the partial decay widths and the production cross sec- luminosity of 6 fb−1 [5]. tion for associatedHiggs production and weak boson fu- 2 v The (3σ) excesses measured by both Atlas [2, 6] sion. Essentially this limits the total Higgs production 9 and CmOs [3, 7] in the h γγ channel might be the first cross section from below (interference effects are typi- 1 → cally small [12, 13]). Gluon fusion, gg h, is sensitive glimpseofalightHiggs(ataconsiderablylargerthanex- → 7 to the propagating heavy fermionic degrees of freedom pected production cross section) which would be in per- 1 and the production rate is again fixed as a function of . fect agreement with all we expect from electroweak pre- 1 the Higgs-fermion couplings. Any new physics extension cision constraints and measurements performed at LEP 1 ormodificationoftheHiggssectorhastoreproducethese andattheTevatron[9]. However,thismightalsobejust 1 unitarityrestoringfeaturestoleaveatheoreticallysound a fluctuation and the Higgs may yet again escape. 1 andpredictivetheoryintheLHC-accessibleenergyrange : What are the implications of these first couple of in- v of .3 TeV in the weak boson fusion channel [14], which verse femtobarns of data for more involved (and better i directly accesses longitudinal gauge boson scattering. In X motivated) scenarios of symmetry breaking? Atlas and extendedHiggssectorsthisistypicallyachievedbylinear r Cmshavealsopresentedstringentboundsonnewphysics a mixing of the various scalar fields hi such that for ener- such as, e.g., the minimal supersymmetric extension of giesmuchlargerthanm thecoherentsumoftheh ex- the SM and the SM with the addition of fourth gener- hi i change diagrams reproduces the SM Higgs contribution. ation. And we expect more model-specific analyses to Thisisthecasefore.g. thetwoHiggsdoubletmodel,the appear soon, when more data becomes available. next to minimal supersymmetric Standard Model [15], Both Atlas and Cms provide strong constraints on composite Higgs models [16] or the Higgs portal scenar- SM-like Higgs production, even if it is suppressed com- ios of Refs. [17, 18]. As a result the Higgs production paredto the SM.Theseconstraintsareexpressedaslim- rates (or more precisely the production rate of the light its on σ/σSM and in some mass ranges σ/σSM 0.2 is SM-like Higgs) can decrease with a characteristic mix- ≃ already excluded at the 95% confidence level [2, 3, 6–8]. ing angle. This can also protect precision electroweak If a Higgs is to exist in this energy range it has to some- observables such as S,T,U [19] from sizable corrections ∗Electronicaddress: [email protected] ∗Unitarization of VLVL → VLVL is a direct consequence of spon- †Electronicaddress: [email protected] taneous symmetry breaking [10], while unitarity in qq¯ → VLVL ‡Electronicaddress: [email protected] relates the fermion and gauge sectors and is less obvious, see e.g. §Electronicaddress: [email protected] Ref.[11]. 2 fromhighmassscales. However,unlessweappealtofine- large SM backgrounds (Sect. IIC). While this cannot be tuning, accounting for both electroweak precision data achieved by changing the gluon coupling alone we find and unitarity at the same time typically amounts to a that it is possible in scenarios with enhanced couplings light SM-like Higgs boson with significant production to light quarks or when heavy flavor mesons decaying cross section [20, 21]. Finally, smaller Higgs production into gluons dominate the Higgs decay chain. We briefly crosssectionscanalsobeobtainedfromanomalousHiggs discuss combinations of these scenarios in Sect. IID. We couplings [22]. conclude this paper with a summary in section III. Themostminimalassumptionisthatelectroweaksym- metry breaking is caused by a single SU(2) doublet Higgs. In this paper we will use this assumption and II. PHENOMENOLOGICAL HIGGS HIDE-OUTS therefore we will focus on invisible decays and modified signatures. In the following we investigate how a sim- ToachieveasituationinwhichtheHiggsphenomenol- ple extension of the Higgs sector can lead to a “hidden” ogy can be hidden we have to introduce an extension of Higgs phenomenology at the LHC. Generically, the re- theSMHiggssectorwhichpreservesSU(2) U(1)gauge cent LHC bounds can be weakened or even avoided this × invariance. We limit ourselvesto renormalizableinterac- way. tionsinthe Higgssector. The onlychoiceisthe addition Such Higgs “hide-out” scenarios are due to a com- ofascalarinteractingviathepreviouslymentionedHiggs bination of dynamics and kinematics, e.g. they occur portal through modified branching ratios as a consequence of an extended spectrum or modified couplings. There is = +η H 2 φ2+∂ φ⋆∂µφ m2 φ2. (1) a plethora of theoretically sound models which do such L LSM | | | | µ − | | modifications and hide the Higgs [18, 21, 23–25]. There- fore,aclassificationonthe levelofthe phenomenological ThefieldφistakentobeasingletundertheSMgauge outcome is desirable. In the following we will pursue group. One could have non-trivial representations of φ this approach. Nevertheless, we also provide an inter- under SU(2)L U(1)Y in the Higgs sector, which ad- × pretation in terms of a simple effective model. Our cat- mitsmoreinvolveddynamicsthanEq.(1). Iftheseextra egorization is of course “non-invertible”; many different degrees of freedom are phenomenologically hidden, we models [26] exhibit a similar phenomenology, and we do encounterasituationwhichstillcanbe meaningfullyde- not try to compile an exhaustive list of model-building scribed by Eq. (1). Throughout, we define the field H realizations of a specific phenomenological outcome. A to be responsible for electroweak symmetry breaking. φ minimal realizationis the coupling of the otherwise SM- can live in the scalar or vectorial representation of the like Higgs to a hidden sector via a renormalizable “por- Lorentz group but can in principle also effectively arise tal” interaction† H 2 [18]. Although it is one of fromafermioniccondensateofastronglyinteractingsec- hid the simplest gaug∼e |inv|arOiant and renormalizable exten- tor. As an additional simplification one can impose a sions we can come up with to model a Higgs hide-out, U(1) or 2 symmetry which forbids terms φ. Z ∼ it is an example of how to evade the currently existing In general there could also be more than one field φ. bounds on SM Higgs and encompasses a huge range of This canallowfor a varietyofdecaycascadesin the hid- phenomenological characteristics. den sector. In our study we will concentrate on the fol- StartingfromourassumptionofanessentiallySM-like lowing simple set of interactions, electroweak symmetry breaking mechanism, our aim is n to categorize the variety of Higgs hide-outs in terms of = ∂ φ⋆∂µφ m2φ⋆φ +∂ φ′⋆∂µφ′ m′2φ′⋆φ′ their phenomenologicalsignatures. At the same time we Lmulti µ i i− i i i µ i i− i i i Xi=1 want to provide a simple parametrizationin terms of an n−1 effective Lagrangian that realizes these features. In sec- + ρ φ⋆φ φ′ +h.c.. (2) tionIIwereviewandcollectthenecessaryingredientsfor i i i+1 i+1 Xi=1 the Lagrangian. In Sect. IIA we consider a dominantly invisibly decaying Higgs and study the implications of IfdesiredonecanchoosechargessuchthattheU(1)sym- currentLHC data. Thenin Sect. IIB we look ata Higgs metryispreserved. Wetaketheparticlestobeorderedin decaying into long lived particles that could be searched mass, allowing cascade decays. We identify the heaviest for by displaced vertex searches. Here we provide ad- state φ with φ in Eq. (1). Also we will typically allow 1 ditional motivation to also search in the outer parts of only the last particles φ of the decay cascade to decay n the detector. Finally the Higgs could also be buried in into SM particles accordingly. If a cascade is considered one has to replace φ φ in Eq. (5) below. n → After the Higgs acquires a vacuum expectation value weinduce aBSMtrilinearcouplingofthe physicalHiggs †There are two more such portals: kinetic mixing with an extra h to φ U(1) gauge group (see e.g. Ref. [27] and Ref. [28] for a review of somelow energy consequences), and neutrinos mixingwith sterile neutrinos(cf.[29]). ǫhφ⋆φ with ǫ=η H =ηv/√2 (3) h i 3 which modifies the Higgs branching ratios for the Higgs decay. In addition the physical mass m is given by 4π φ 10 m2 =m2 η H 2 =m2 ηv2/2, (4) η φ − h i − g n which we take to be positive. Note that since φ does not pli develop a vacuum expectation value, there is no mixing ou 1 c of the two scalar states. Accordingly cross sections and al decay rates are not modified by mixing effects. rt o If unbroken, the global U(1) symmetry forbids decays p of φ into SM particles. Hence, in order to re-introduce 0.1 suchdecaysweneedU(1)breakingcouplingsto SMpar- Γinv/Γtot =0.9 ticles. The SMgaugesymmetriesforbidcouplingsofSM 100 200 300 400 500 matter and gaugefields to φ on the renormalizablelevel. mh [GeV] Wewillthereforeconsiderthefollowingdimension5cou- plings(whichautomaticallyalsoexplicitlybreaktheU(1) FIG.1: RequiredsizeoftheHiggsportalcouplingηtoachieve symmetry), an invisible branching ratio of BR(invis) = 90%. We choose m = 10 GeV, the result is however rather independent of φ λij this choice. φD HΨ + h.c. L,i R,j L⊃ M Xi κ κ˜ Eq. (5) vanish or are very small, φ is stable with respect + γ(φ+φ⋆)FµνF + γ(φ φ⋆)FµνF˜ (cid:20)M µν M − µν(cid:21) to decays into Standard Model particles. Accordingly, the decay h φ⋆φ (possible as long as m 2m ) is + κg(φ+φ⋆)GµνG + κ˜g(φ φ⋆)GµνG˜ (5) invisible. → h ≥ φ µν µν (cid:20)M M − (cid:21) Such invisible decays make search strategies based on where D denotes the left-handed fermion doublet and visibleSMparticlesmoredifficult. Naively,onecanagain L Ψ the right handed fermions. In our effective theory use that current Higgs searches are already sensitive to R approachwe can choose suitable λ to avoidflavor, lep- cross sections lower than the SM cross section. Using ij ton number and baryonnumber changing processes [30]. this one can reinterpret the limits on σh/σhSM as limits F˜, G˜ arethe dualQEDandQCDfield strengthtensor‡. on σhBR(h→SM)/σhSM. Thesehigherdimensionaloperatorsallowφtodecayback The partial decay width for h φ⋆φ is given by → to visible SM matter. η2v2[(m2 4m2)]1/2 Since we take the Higgs to be responsible for elec- Γ = h− φ . (8) troweak symmetry breaking the partial decay widths for inv 16π m2h h VV are fixed. However, we can model modifica- → As can be seen from Fig. 1 at low Higgs masses quite tions of the branching ratios and the total decay width moderatevaluesofη aresufficientto achievedominantly byintroducinganadditionalcontributiontothecoupling invisibledecays,allowingtoevadesearchstrategiesbased =χ αs hGµνG , (6) on SM particles for now. Even when we later on look at Lggh 12πv µν modified decays to SM particles this makes it more diffi- culttohidetheHiggsinthehighmassregion. Abovethe (see Ref. [31] for theoretical bounds) or operators of the VV-threshold, however, the hidden sector decay width form has to compete with a much larger and rapidly grow- LHHqq = βMQL2H†HQ¯LD/QL. (7) finaigrl∼y lmar3hgedveaclauyeswoidftthheinHtoigSgsMpopratratliccloesu.plTinhgisηr&equ1irfoers m &180GeVandevenη &4π form &450GeV. The Ofcourseonecanalsoaddsimilarcouplingsfortheright h h reason for this is as simple as compelling: Unitarization handed quarks and leptons. of VV scattering requires a sufficiently large coupling of the Higgs to energetic longitudinal Vs resulting also in A. Hidden Higgses a large particle decay width of the Higgs into those if the Higgs is heavy. The required large couplings to hide the Higgs into invisible decays are at odds with pertur- If the global U(1) symmetry of the φ-field is unbroken bativity unless they are connected to spontaneous sym- or extremely weakly broken, i.e. the couplings given in metry breaking. Therefore, at large Higgs masses hiding the Higgs into invisible decays is difficult from a model building point of view. For low Higgs masses the Higgs can be easily hidden in invisible decays and alternative ‡Ifkinematicallypossible,i.e. forveryhighHiggsmasses,onecould also add an additional term for decays to W±,Z. This however strategies[33,34]basedonmissingenergy(E/T)searches wouldnotleadtoaphenomenologically verydifferentsituation. then become necessary. 4 100 −1 rithm [41], with R=0.4, and we keep only events where L=1 fb L C pj1 >120 GeV, ηj1 <2 % T | | 95 pj2 <30 GeV if ηj2 <4.5 (jet-veto) T | | M E/ >120 GeV. (9) Ss)/σh 10 LowaPpTprmoxo.nosijgent+alE/eTfficciheanncnieesl We obtain E/TT as the transverse momentum of the total vi momentum from all the visible particles, defined to be n ±2σ edge (i final state particles with p > 2 GeV and η < 4.5. We R ±1σ edge T | | B alsorejectedthe (rare)eventswhere there is at leastone h expected central 95%CL σ 1 observed central 95%CL lepton within the cuts defined in Ref. [37]. We extract the background distribution and the data 100 150 200 250 300 350 400 450 500 for a luminosity =1 fb−1 from Ref. [37]. mh [GeV] In Fig. 2 we pLlot the resulting 95% confidence level (CL) upper bound on σ BR(invis)/σSM applying the FIG. 2: Expected and observed (hξi = 0.27) 95% CL upper CL method¶. h h S limits on the production cross section in multiples of theSM We estimate the averageefficiency ξ that relates the cross section for a monojet+E/T search confronted with an theoreticalHiggscrosssectionpredictihoniforagivenmass invisibly decaying Higgs. to the experimentally observable one by performing a Monte Carlo study relating the results of the dominant and signature-wise similar (Z inv)+jets background An important constraint for hidden Higgses comes → to the expected numbers quoted in Ref. [37]. For this from direct searches of associated hidden Higgs produc- purpose we produce a (Z inv)+jets event sample us- tion at LEP. For SM-like Higgs gauge boson couplings ing Sherpa [44] and norm→alize it to the next-to-leading this puts a lower limit of 114.4 GeV on the mass of a order QCD cross section obtained with Mcfm [45] dominantly invisibly decaying Higgs [35]. (σNLO (pj 120 GeV, η 2.0)= 41.3 pb). Run- (Z→ν¯ν)+j T ≥ | j| ≤ ning the analysis with the above cuts we can estimate ξ = 0.27 and we use this significance to rescale our Monojet + missing energy channel h i Higgs signal hypotheses in Fig. 2. The observed limits are weaker than the expected limits because of a slight At the LHC the production of a light Higgs boson is excess of the central data values [37]. dominated by the gluon fusion process gg h+X. Ac- Theconfidencelevelscalesas −1/2. Whilesearches → cordingly a search in the monojet plus missing energy ∼L with present luminosities are insensitive to SM-like pro- channel is promising [36] because backgrounds are com- duction cross sections (χ 1 in Eq. (6)) enhanced pro- parably small and can be brought under sufficient con- ≃ ductioncrosssections(χ 3 5)occurringinavarietyof trol [37]. We focus in the following on a center of mass ≃ − models are already constrained. In particular for fourth energyof7TeV,whichallowstorelateourresultstothe generation models [46–49] where we expect χ 3, this current LHC run. ≃ will very soon become a relevant constraint. With the We compute the gg h + X signal using an NLO fast growing data sets these constraints will tighten sig- → computation matched to a parton shower§: the inclu- nificantly in the near future. sive cross section is therefore NLO accurate, the hardest jet (relevant for this study) is described with the full h+1 jet matrix element accuracy and further emissions Two Leptons + E/ Channel T are generated in the shower approximation. We use the Powheg method (as implemented in the Powheg Box A very clean and therefore important search chan- program) [39], together with (transverse-momentum or- nel for hidden Higgs decays is associated production dered) Pythia 6 [40]. We generate events for values of pp hZ with subsequent decay of the Z to leptons m between 100 and 500 GeV, using the narrow-width → h [33, 50]. Note that for associated production, it is much approximation. moredifficultto obtainincreasedcrosssectionssince the The Higgs-boson is set stable and excluded from the relevantcouplingisfixedbygaugeinvariancek. Forcom- tracks entering the analysis. On all the other final state pleteness and comparison with the previous section, we particles we apply the “LowPT” selection cuts described in Ref. [37]. Jets are constructed using the anti-k algo- t ¶The CLS method procedure is fairly standard method in nowa- days experimental analysis to present constraints in new physics §Thematchingprescriptioninthischannelissubjecttoanongoing searches. WereferthereadtoRefs.[42,43]forfurtherdetails. discussion in the corresponding community, see Refs. [38, 39] for kThis fact also accounts for the hidden higgs search in the weak details. Ourresultsdonotincludeanytheoretical uncertainties. boson fusion channel discussed in [51]. For a center of mass en- 5 100 ±2σ edge center of mass energy. Eventually this channel will be- L ±1σ edge come againimportantfor largeluminosities (100fb−1) C O % expected 95%CL at √s=14 TeV [33]. 5 9 M Sh 10 B. Reemerging Higgses σ / ) s nvi In some cases the “invisible” decay products of the R(i two leptons+ E/T channel previoussubsectioncandecaybackintoSMparticlesand B noefficiencies, nosystematics thehiddenHiggsslowlyreemerges. Inourtoymodelthis h σ 1 is realized when the U(1) violating couplings of Eq. (5) are turned on. At leading order the decay rate of φ is 100 120 140 160 180 200 given by, mh [GeV] Y2(m2 4m2)3/2 FIG.3: Expected95%CLupperlimitsontheassociatedpro- Γ = φ− f (10) φ 4π 2m2 ductioncrosssectioninmultiplesoftheSMcrosssectionfora φ twoleptons+E/ searchconfrontedwithaninvisiblydecaying T Higgs. Superficially this channel looks more sensitive then where Y = λv/(√2M) is the effective Yukawa coupling the monojet search. Note, however, that this plot does not thatresultsfromthe lagrangianEq.(5). Ifthe couplings include any efficiencies and is not based on actual data. are very small, i.e. the mass scale M is very high, φ can travel a measurable distance before decaying, βγ nonetheless estimate the performance of the correspond- d= . (11) Γ ing search. To our knowledge there is no publicly avail- φ ableLHCresultoftheprocessesinthephasespaceregion Inthiscaseonemaysearchfor(highly)displacedvertices weareinterestedin. Hence the results ofthis sectionare [54] or use adapted trigger strategies [55, 56]. If decays obtainedfromMonteCarlo,andweincludeneitherback- happen inside the trackerthis is a fairly clean signature. ground nor signal systematics (this includes a potential If the decay length is of the order of meters and above mismeasurement of Z+jets, giving rise to a finite E/T by a significantpart of the decayswill happen in the bigger detector effects). Therefore, the sensitivity in this chan- outerpartsofthedetector(orevenoutsidethedetector). nel is obviously optimistic. In this case one can gain additional sensitivity by also The cross section for associated Higgs production is triggering on events where the decay happens outside of much smaller than the one for gluon fusion, and shape the tracker. If the decay happens outside the detector comparisons of e.g. the pZ distribution are not possible coverage the same search strategies outlined in Sec. IIA T given .10 signal events for =1 fb−1. Instead we per- apply. form a counting experimentLfor the signal and dominant TheLHCexperimentsAtlasandCmsconsisteachof ZZ, WW and tt¯backgrounds. Again we compute the four different layers: the inner tracker, the electromag- signal and background cross sections at √s=7 TeV us- netic(E)calorimeter,thehadron(H)calorimeterandthe ing Mcfm. We require two oppositely charged leptons of identical flavor to combine to the Z mass within a ±10 GeV mass window and E/T ≥100 GeV. 1 hhddii≃≤41mm Our resulting estimate on the upper 95% CL is shown hdi≃7m inFig.3. Forverylowmassesthetwoleptons+E/ chan- 0.8 T hdi≃20m nel can be of similar importance as the monojet+E/ T y t search,depending on the signaland backgroundefficien- bili 0.6 cies. For higher Higgs masses this channel looses sensi- a b tivityveryquicklyduetothesmallcrosssectionat7TeV o pr 0.4 0.2 ergyof7TeVweakbosonfusionisnotanimportantchannelsince for typical search cuts and L ≃ 1fb−1. We find a reduced cross 0 section [52] σ(7 TeV)/σ(14 TeV) ≃ 0.2 compared to 14 TeV. As tracker E/H µdetec. outside a consequence searches based on small angular separations of the two tagging jets are less sensitive compared to other channels if FIG.4: Probability,basedontheAtlasgeometry,foraφto onealsotakesintoaccountthesystematicuncertaintiesofthecen- decay in the tracker, E/H-calorimeter, muon-calorimeter or tral jet veto and the forwardtag jet energy scale uncertainty due to pile-up. Only recently, with the 5 fb−1 set Cms has started to outside the detector. We show the results for four different average lab frame decay lengths, respectively. overcomethesesystematiclimitations[53]. 6 transverse momentum (i.e. when the mass difference to the sum of the masses of the decay products is small) and the decay lengths are all equal λ/n the resulting probability distribution for an n-step decay is y bilit exp( nx/λ)n(nx)n−1λ−n ba Pn(x)= − . (13) o (n 1)! pr − This is shown in Fig. 5. In this way one can reduce the number of decays happening in the inner tracker com- pared to decays occurring in outer parts increasing the 0 need to also check for decays there. track. E+Hcal. muon outside FIG. 5: Probability for the distance between initial Higgs C. Buried Higgses production and final decay back into SM particles for a cas- cade decay with 1 (blue), 2 (red), 5 (yellow) and 50 (green) NaivelywecanalsotrytohidetheHiggsbyburyingit intermediate steps. inthe busy hadronicfinalstate atthe LHC.Inoureffec- tive theory approach we can facilitate this by increasing its branching ratio to gluons via the operator of Eq. (6). muon calorimeter. The sensitivity on highly displaced The modified decay width to gluons then scales like vertices is limited by the radial extension of the experi- ments. The radial size of the different detector compo- Γ(h gg)=χ2Γ (h gg). (14) SM → → nents differs between the two experiments, particularly for the muon calorimeters. We focus in the following The same effect increasesthe higgsproductioncrosssec- analysisoptimistically onthe geometryofthe largerAt- tionfromgluonfusion by the same factor,yielding mod- las experiment. ified production times branching ratio factors for the in- The muon calorimeter of the Atlas experiment is dividual higgs decay channels gg h ii(†) → → located d . 11 m away from the interaction point. [σBR] χ2 The muon detector coverage in pseudorapidity is η < i=g : = , 2.4. In the muon calorimeter photons and elec|tr|ons 6 [σBR]SM (χ2 1)BRSM(h gg)+1 − → (15) will be stripped in an early layer after the conversion [σBR] χ4 φ ff¯, hence, are likely to be misinterpreted as detec- i=g : [σBR] = (χ2 1)BR (h gg)+1. → SM SM tornoise[57]. AssumingtheHiggsbosondecaysinstanta- − → neously the probability for φ to decay between distances For the “standard” search channels we have d and d >d is given by 1 2 1 [σBR] 1 for χ 1, (16) d2 Γ Γx [σBR]SM ≫ ≫ p(d d d )= dx exp . (12) 1 2 ≤ ≤ Z βγ (cid:18)−βγ(cid:19) d1 since for light Higgs masses we have BRSM(h gg) = → (1%). AsaresulttheHiggsisnotburiedbutevenmore In the lab frame the decay of h φ⋆φ induces a sig- O → visible in the standard search channels. Note, however, nificant boost factor β for φ. We find, that the trans- that the increased total width of the Higgs boson will verse momentum of the Higgs boson generated from ini- giverisetoreducedreconstructionefficienciesinstandard tialstateradiationis ofminorimportanceforβ, because search channels∗∗. the Higgs rarely decays along its direction of motion. One way to bury the Higgs in hadronic final states is Based on Eq. (12) we show the probability of φ decay- to enhance one of the couplings to light quarks, for ex- ing respectively in the tracker, the E/H-calorimeter, the ample using the operator of Eq. (7). This reduces the muon-calorimeter or outside the detector in Fig. 4 for a branchingratiosforthestandardsearchsignatureswith- number of decay lengths. We include the effects of finite out significantly increasing the production cross section, detector coverage η 2.4 for a particle cluster with | | ≤ sincegluonfusiondominatestheproduction. Thismight p 2 GeV and correct on the longitudinal dimension T ≥ be one of the most difficult channels to uncover. of the detector. Alternativelywecanagainconsiderdecaystoφswhich So far the larger number of decays happening in the promptlydecayandbackintoSMquarksorgluons. This outerpartsofthedetectorarisesimplyfromthefactthat these parts are bigger. However, if we allow for cascade decays h φ φ ... SM as described by the 1 2 → → → → Lagrangian Eq. (2) one can achieve that more decays ∗∗In principle a similar effect can arise if we dramatically increase happen close to the average total decay length. For the an invisible decay width. The monojet and other invisible Higgs simple case when subsequent decays generate negligible searchstrategiesshouldbelessaffected bythis. 7 1.4 mφ≤10.4 GeV subscriptindicatesthedecaypropertiesoftheφasinvis- m ≃10.6 GeV ible, long-lived or dominant decay into jets. φ 1.2 m ≥11.0 GeV φ h φ (φ ,φ ): Adapted trigger strategies to 1 → disp inv jets y highlydisplacedverticesasoutlinedin[55]canobviously t bili 0.8 mh=130 GeV cope with signatures of these types as well, unless the a b majorityofdecaysocuroutside the detector. Eventually o 0.6 pr reconstructing the mass of the Higgs boson is however 0.4 more involved. This is due to the systematic uncertain- ties that enter in the measurement of the E/ , which is 0.2 T sensitive to all energy deposited in the detector, or the 0 jet energy scale. 0 1 2 3 4 numberof long-lived B hadrons h φinvφjets: If the Higgs is produced at rest and → m m the resulting signature is a monojet with a h ≫ φjets FIG. 6: We count the number of long lived B-hadrons with missing energy mh/2. In this case boosting the Higgs ∼ decaylengthcτ ≥0.3mmforahiggsdecayingviah→φφ→ by recoiling it against a Z or an additional jet is not b¯bb¯b. necessarily a good strategy. Then the jet from φ and jets the missing energy from φ are aligned. This is prob- inv lematic since it is already difficult to precisely measure hides the Higgs not only by decaying into jets but these the jet energy. Therefore it is difficult to measure miss- jets are also softer. Examples of these scenarios are re- ing energy aligned with jets. In fact, Atlas and Cms alized in Refs. [25, 58, 59]. Introducing cascades as in typically apply cuts on extra jets in E/ searches to re- T Eq.(2) producesmore butsofter radiation. Thereforeto movebackgrounds,reducingthe sensitivity towardthese identify the Higgs decay products it is advantageous to finalstates[63]. Z+jets is alreadya non-negligibleback- produce the Higgs in a boosted state. Irrespectiveof the ground if the Higgs recoils against a Z but this problem presenceofacascadethis canhelpdisentanglethe Higgs becomes even more severe if the Higgs recoils against a from the background [59, 60]. This also helps when one jet. uses a cascade ending in partially or completely leptonic finalstateswhichwouldotherwisebetoosofttopassthe cuts [61]. III. SUMMARY Another option is to decay the Higgs into gluons via a φ with a mass close to a heavy flavor bound state, In this paper we have discussed phenomenological e.g. b¯b. In the φ rest frame the quarks are back to back modificationsofthe Higgssectorwhichcanserveto hide and if their momenta are small it is likely for them to a Higgs and weaken the currently existing bounds. The hadronize to a quark-anti quark bound state. Quark- Standard Model-like Higgs is a minimal solution to elec- anti quark bound states like Υ and η typically decay b troweak symmetry breaking. Therefore we focus on sce- dominantly into gluons. In this case no direct coupling narioswheretheelectroweaksymmetrybreakingsectoris of the φ to gluons and light quarks is necessary and this asintheStandardModel(SM).Thislimitsthenumberof wouldbeagenericfeatureiftheφcouplesYukawa-liketo potentialhide-outsfortheHiggsbosonandallowstoclas- quarks. One example is a cascade h φφ ΥΥ 6g. sifythemintermsofphenomenologicalsignaturesatcol- In Fig. 6 we use Pythia 8 [62] to s→imulat→e such →a cas- liders. We also suggest a simple benchmark model that cade. This example demonstrates that the φ needs to parameterizesthese possible signatures. Modified search be quite close to threshold as already a relatively small strategies, looking at different signatures but also using mass difference between the φ and the quark-anti quark adapted trigger strategies and elaborate reconstruction bound state significantly reduces the branching ratio to techniques (e.g. subjets) can help to uncover even an this bound state leaving us with a significant number of evasive Higgs. long lived B-hadrons. For other “heavy” quarks identi- There are essentially three possible hide-outs. One is fication is more difficult to begin with as there are fewer to have dominantly invisible decays. This typically pro- clean long-lived states. duces missing energy signatures. In this case a promis- ing search channel is a monojet plus missing energy. We have used existing Atlas data for this channel and D. Combined Higgs hide-outs constrained the production cross section times invisible branching fraction. For comparison we also estimated The hide-outs outlined in the previous sections IIA- the sensitivity for searches using associated production. IIC can be combined and result in a large variety of Asecondhide outis forthe Higgsistofirstdecayinto signatures some of which can be even more challenging. long-lived neutral particles. Their eventual decay may Let us imagine that the Higgs can decay via three dif- then be observed in displaced vertex searches. Depend- ferent states denoted by φ , φ , and φ , where the ing on the decay length it may be advantageous to also inv disp jets 8 searchforsuchdecaysoccurringinthe outerpartsofthe straints for all hide-outs in the near future. detector. Thisfeaturecouldevenbeenhancedbycascade decays. Finally, the Higgs could decay dominantly to light quarks and gluons and be buried in huge QCD back- Acknowledgements ground. These decays canoccur via cascades potentially includingalsointermediateq¯q boundstates. Here,meth- odslookingintospecificradiationpatternsinthejetsub- We thank Kristov Hackstein, Marc Hohlfeld, Peter structure can be important to uncover the Higgs. Loch, Andrew Pilkington and David Strom for helpful Of course,combining the different hide-outs can make discussions. CE acknowledges funding by the Durham searches even more difficult. International Junior Research Fellowship scheme. ER In general we find that hiding heavy Higgses is rather acknowledges financial support from the LHCPhenoNet network under the Grant Agreement PITN-GA-2010- challenging from a theoretical point of view while light Higgses (for which h VV is inaccessible) is more 264564. → straightforward. The simulations underlying this study have been per- Current analysis are not yet sensitive enough to pin formed in parts on bwGRiD, member of the German down all Higgs hide-outs. However, the ever increas- D-Grid initiative, funded by the Bundesministerium fu¨r ing LHC data set together with a combined analysis of Bildung und Forschung and the Ministerium fu¨r Wis- all sensitive channels will significantly tighten the con- senschaft, Forschung und Kunst Baden-Wu¨rttemberg. [1] G. Aad et al. [Atlas Collaboration], JINST 3 (2008) misek and J. F. Gunion, Phys. Rev. Lett. 95 (2005) S08003. G. L. Bayatian et al. [CMS Collaboration], J. 041801. U. Ellwanger, J. F. Gunion and C. Hugonie, Phys.G 34 (2007) 995. JHEP 0507 (2005) 041. G. G. Ross and K. Schmidt- [2] The Atlas Collaboration, arXiv:1112.2577 [hep-ex] Hoberg, arXiv:1108.1284 [hep-ph]. ATLAS-CONF-2011-161, ATLAS-CONF-2011-162, [16] B. Gripaios, A. Pomarol, F. Riva and J. Serra, JHEP ATLAS-CONF-2011-163. 0904 (2009) 070. [3] The CMS Collaboration, CMS-PAS-HIG-11-025, CMS- [17] M.J.StrasslerandK.M.Zurek,Phys.Lett.B651,374 PAS-HIG-11-030, CMS-PAS-HIG-11-031, CMS-PAS- (2007). HIG-11-032. [18] R. Schabinger and J. D. Wells, Phys. Rev. D 72 (2005) [4] R.Barate et al.,Phys. Lett. B 565 (2003) 61. 093007. R. Foot, H. Lew and R. R. Volkas, Phys. [5] The Atlascollaboration, ATL-PHYS-PUB-2011-001. Lett. B 272 (1991) 67. Z. Chacko, H. S. Goh and [6] The Atlas collaboration, ATLAS-CONF-2011-134, R. Harnik, Phys. Rev. Lett. 96 (2006) 231802. B. Patt ATLAS-CONF-2011-111. and F. Wilczek, arXiv:hep-ph/0605188. V. Barger, [7] The CMS collaboration, CMS-HIG-11-014, CMS-PAS- P. Langacker, M. McCaskey, M. J. Ramsey-Musolf and HIG-11-003. G. Shaughnessy, Phys. Rev. D 77, 035005 (2008). [8] The Atlas collaboration, ATLAS-CONF-2011-157. The T. Binoth and J. J. van der Bij, Z. Phys. C 75, 17 CMS collaboration, CMS PAS HIG-11-023. (1997). O. Bahat-Treidel, Y. Grossman and Y. Rozen, [9] The ALEPH, CDF, D0, DELPHI, L3, OPAL, SLD Col- JHEP 0705, 022 (2007). A. Dedes, T. Figy, S. Hoche, laborations, the LEP Electroweak Working Group, the F. Krauss and T. E. J. Underwood, JHEP 0811 (2008) TevatronElectroweakWorkingGroup,andtheSLDelec- 036. troweakandheavyflavourgroups,arXiv:1012.2367 [hep- [19] M. E. Peskin and T. Takeuchi, Phys. Rev. D 46, 381 ex]. (1992). [10] J. M. Cornwall, D. N. Levin and G. Tiktopoulos, Phys. [20] R. Barbieri, A. Pomarol, R. Rattazzi and A. Strumia, Rev.Lett.30(1973)1268[Erratum-ibid.31(1973)572]. Nucl. Phys. B 703, 127 (2004). G. F. Giudice, C. Gro- J. M. Cornwall, D. N. Levin and G. Tiktopoulos, Phys. jean, A. Pomarol and R. Rattazzi, JHEP 0706 (2007) Rev.D10(1974) 1145 [Erratum-ibid.D 11(1975) 972]. 045.K.AgasheandR.Contino,Nucl.Phys.B742(2006) [11] K.Hagiwara,R.D.Peccei,D.ZeppenfeldandK.Hikasa, 59. S. Bock, R. Lafaye, T. Plehn, M. Rauch, D. Zerwas Nucl.Phys. B 282, 253 (1987). and P. M. Zerwas, Phys.Lett. B 694 (2010) 44. [12] A.Bredenstein,K.HagiwaraandB.Jager,Phys.Rev.D [21] C.Englert,T.Plehn,D.ZerwasandP.M.Zerwas,Phys. 77 (2008) 073004. Lett. B 703 (2011) 298. [13] M. Ciccolini, A. Denner, S. Dittmaier, Phys. Rev. [22] K.Hagiwara,R.SzalapskiandD.Zeppenfeld,Phys.Lett. Lett. 99 (2007) 161803. M. Ciccolini, A. Denner and B318(1993)155.V.Hankele,G.Klamke,D.Zeppenfeld S.Dittmaier, Phys.Rev. D 77 (2008) 013002. andT.Figy,Phys.Rev.D74(2006) 095001. F.Bonnet, [14] J. Bagger et al.,Phys.Rev.D 52 (1995) 3878. B. Jager, M. B. Gavela, T. Ota and W. Winter, arXiv:1105.5140 C. Oleari and D. Zeppenfeld, Phys. Rev. D 73 (2006) [hep-ph]. 113006. B. Jager, C. Oleari and D. Zeppenfeld, JHEP [23] W. D. Goldberger, B. Grinstein and W. Skiba, Phys. 0607 (2006) 015. Rev.Lett.100(2008) 111802 J.Fan,W.D.Goldberger, [15] U. Ellwanger, M. Rausch de Traubenberg and A. Ross and W. Skiba, Phys.Rev.D 79 (2009) 035017 C. A. Savoy, Nucl. Phys. B 492 (1997) 21. R. Der- [24] Y. Mambrini, arXiv:1108.0671 [hep-ph]. M. Raidal, 9 A. Strumia, arXiv:1108.4903 [hep-ph]. R. Foot, http://fastjet.fr/ A. Kobakhidze and R. R. Volkas, arXiv:1109.0919 [hep- [42] A.L.Read,CERN-OPEN-2000-205.A.L.Read,J.Phys. ph]. X. -G. He, J. Tandean, Phys. Rev. D84, 075018 G G28 (2002) 2693-2704. (2011). M. Pospelov, A. Ritz, arXiv:1109.4872 [hep-ph]. [43] T. Junk, Nucl. Instrum. Meth. A 434 I. Low, P. Schwaller, G. Shaughnessy,C. E. M. Wagner, (1999) 435. T. Junk, CDF Note 8128 arXiv:1110.4405 [hep-ph]. E. Weihs and J. Zurita, [cdf/doc/statistics/public/8128]. T. Junk, CDF Note arXiv:1110.5909 [hep-ph]. J. W. Cui, H. J. He, L. C. Lv 7904 [cdf/doc/statistics/public/7904]. H. Hu and J. andF.R.Yin,arXiv:1110.6893 [hep-ph].S.V.Demidov, Nielsen, in 1st Workshop on Confidence Limits, CERN D. S. Gorbunov and A. A. Tokareva, arXiv:1111.1072 2000-005 (2000) [hep-ph]. [44] T. Gleisberg, S. Hoeche, F. Krauss, M. Schonherr, [25] B. Bellazzini, C. Csaki, A. Falkowski, A. Weiler, Phys. S. Schumann, F. Siegert and J. Winter, JHEP 0902 Rev. D81 (2010) 075017. B. Bellazzini, C. Csaki, (2009)007;S.Schumann,F.Krauss,JHEP0803 (2008) A.Falkowski,A.Weiler,Phys.Rev.D80 (2009)075008. 038; S. Hoeche, F. Krauss, S. Schumann, F. Siegert, [26] D.Alves et al.,arXiv:1105.2838 [hep-ph]. JHEP 0905 (2009) 053. [27] L. B. Okun, Sov. Phys. JETP 56 (1982) 502 [Zh. Eksp. [45] J.M.CampbellandR.K.Ellis,Phys.Rev.D60,113006 Teor.Fiz.83(1982)892].B.Holdom,Phys.Lett.B166, (1999). J.M.CampbellandR.K.Ellis, arXiv:1007.3492 196 (1986). [hep-ph].http://mcfm.fnal.gov. [28] J. Jaeckel and A. Ringwald, Ann. Rev. Nucl. Part. Sci. [46] G.D.Kribs,T.Plehn,M.SpannowskyandT.M.P.Tait, 60 (2010) 405. Phys. Rev.D 76 (2007) 075016. [29] S. Dodelson and L. M. Widrow, Phys. Rev. Lett. 72 [47] B.Holdom,W.S.Hou,T.Hurth,M.L.Mangano,S.Sul- (1994)17J.March-Russell,S.M.West,D.Cumberbatch tansoy and G. Unel,PMC Phys.A 3(2009) 4. and D. Hooper, JHEP 0807, 058 (2008). [48] X. Ruan and Z. Zhang, arXiv:1105.1634 [hep-ph]. [30] C.D.Froggatt,H.B.Nielsen,Nucl.Phys.B147 (1979) [49] K. Belotsky, D. Fargion, M. Khlopov, R. Konoplich and 277. K.Shibaev,Phys.Rev.D68(2003)054027.S.A.Cetin, [31] I.Low, R.Rattazzi, A. Vichi,JHEP 1004 (2010) 126. T. Cuhadar-Donszelmann, M. Sahin, S. Sultansoy and [32] V. Barger, P. Langacker, G. Shaughnessy, Phys. Rev. G. Unel, arXiv:1108.4071 [hep-ph]. L. M. Carpenter, D75 (2007) 055013. V. Barger, P. Langacker, M. Mc- arXiv:1110.4895 [hep-ph]. Caskey, M. J. Ramsey-Musolf, G. Shaughnessy, Phys. [50] R. M. Godbole, M. Guchait, K. Mazumdar, S. Moretti Rev. D77 (2008) 035005. V. Barger, P. Langacker, andD.P.Roy,Phys.Lett.B571(2003)184.H.Davoudi- M. McCaskey, M. Ramsey-Musolf, G. Shaughnessy, asl, T. Han and H. E. Logan, Phys. Rev. D 71 (2005) Phys. Rev. D79 (2009) 015018. V. Barger, H. E. Lo- 115007. gan, G. Shaughnessy,Phys. Rev.D79 (2009) 115018. [51] O.J.P.Eboli,D.Zeppenfeld,Phys.Lett.B495,147-154 [33] A.De Roeck et al.,Eur. Phys. J. C 66 (2010) 525. (2000). [34] K. Belotsky, V. A. Khoze, A. D. Martin and [52] K. Arnold et al., Comput. Phys. Commun. 180, 1661 M. G. Ryskin,Eur. Phys.J. C 36 (2004) 503. (2009); K.Arnold et al.,arXiv:1107.4038 [hep-ph]. [35] LEP Higgs Working for Higgs boson searches, [53] The CMS collaboration, CMS-PAS-HIG-11-029. arXiv:hep-ex/0107032. [54] M. J. Strassler, K. M. Zurek, Phys. Lett. B661 (2008) [36] A.Birkedal,K.Matchev,M.Perelstein,Phys.Rev.D70, 263-267. 077701(2004).J.L.Feng,S.SuandF.Takayama,Phys. [55] G. Ciapetti, ATL-COM-PHYS-2008-155. A. Nisati, Rev. Lett. 96, 151802 (2006) M. Beltran, D. Hooper, S. Petrarca, G. Salvini, ATL-MUON-97-205. S. Am- E.W.KolbandZ.C.Krusberg,Phys.Rev.D80,043509 brosanio, B. Mele, S.Petrarca, G.Polesello, A.Rimoldi, (2009)Q.-H.Cao,C.-R.Chen,C.S.Li,H.Zhang,JHEP ATL-PHYS-2002-006.S.Taremetal.,ATL-PHYS-PUB- 1108, 018 (2011). M. Beltran, D. Hooper, E. W. Kolb, 2005.022. J. Ellis, A.R. Raklev, O.K. Oye, ATL-PHYS- Z. A. C. Krusberg, T. M. P. Tait, JHEP 1009, 037 PUB-2007-016. S. Tarem, S. Bressler, H.Nomoto, A. Di (2010). J. Goodman, M. Ibe, A. Rajaraman, W. Shep- Mattia, ATL-PHYS-PUB-2008-01. herd, T. M. P. Tait, H. -B. Yu, Phys. Lett. B695, 185- [56] S. Dimopoulos, M. Dine, S. Raby and S. D. Thomas, 188 (2011). Y. Bai, P. J. Fox, R. Harnik, JHEP 1012, Phys. Rev. Lett. 76 (1996) 3494. C. H. Chen and 048(2010).P.J.Fox,R.Harnik,J.Kopp,Y.Tsai,Phys. J. F. Gunion, Phys.Rev.D 58, 075005 (1998). Rev. D84, 014028 (2011). A. Rajaraman, W. Shepherd, [57] Peter Loch, privatecommunication. T.M.P.Tait,A.M.Wijangco,arXiv:1108.1196[hep-ph]. [58] S. Chang, R. Dermisek, J. F. Gunion and N. Weiner, P. J. Fox, R. Harnik, J. Kopp, Y. Tsai, arXiv:1109.4398 Ann.Rev.Nucl. Part. Sci. 58 (2008) 75. [hep-ph]. [59] A.Falkowski,D.Krohn,L.-T.Wang,J.Shelton,A.Tha- [37] The Atlas collaboration, arXiv:1106.5327 [hep-ex], and lapillil,arXiv:1006.1650[hep-ph].C.R.Chen,M.M.No- ATLAS-CONF-2011-096 for the updated analysis of the jiriandW.Sreethawong,JHEP1011(2010)012.B.Bel- L=1 fb−1 data set. lazzini,C.Csaki,J.HubiszandJ.Shao,Phys.Rev.D83 [38] S. Hoeche, F. Krauss, M. Schonherr and F. Siegert, (2011)095018C.Englert,T.S.RoyandM.Spannowsky, arXiv:1111.1220 [hep-ph]. arXiv:1106.4545 [hep-ph]. [39] P. Nason, JHEP 0411, 040 (2004). S. Alioli, P. Nason, [60] J. M. Butterworth, A. R. Davison, M. Rubin, C. Oleari and E. Re, JHEP 0904, 002 (2009). S. Alioli, G. P. Salam, Phys. Rev. Lett. 100, 242001 (2008). P.Nason, C. Oleari, E. Re, JHEP 1006, 043 (2010). T.Plehn,G.P.Salam,M.Spannowsky,Phys.Rev.Lett. [40] T.Sjostrand, S.Mrenna, P.Z.Skands,JHEP 0605, 026 104, 111801 (2010). G. D. Kribs, A. Martin, T. S. Roy (2006). and M. Spannowsky, Phys. Rev. D 81 (2010) 111501. [41] M. Cacciari, G. P. Salam, G. Soyez, JHEP 0804, D. E. Soper, M. Spannowsky, JHEP 1008, 029 (2010). 063 (2008). M. Cacciari, G.P. Salam and G. Soyez, L.G.Almeida,S.J.Lee,G.Perez,G.Sterman,I.Sung, 10 Phys. Rev. D82, 054034 (2010). G. D. Kribs, A. Mar- 1107, 060 (2011). tin,T.S.Roy,M. Spannowsky,Phys.Rev.D82,095012 [61] C. Cheung, J. T. Ruderman, L. T. Wang and I. Yavin, (2010). C. Hackstein, M. Spannowsky, Phys. Rev. D82, JHEP 1004 (2010) 116. A. Falkowski, J. T. Ruderman, 113012(2010).C.Englert,C.Hackstein,M.Spannowsky, T.VolanskyandJ.Zupan,JHEP1005(2010)077;Phys. Phys. Rev. D82, 114024 (2010). D. E. Soper, M. Span- Rev. Lett.105 (2010) 241801. nowsky, Phys. Rev. D84, 074002 (2011). J. -H. Kim, [62] T. Sjostrand, S. Mrenna, P. Z. Skands, Comput. Phys. Phys. Rev. D83, 011502 (2011). J. Gallicchio, J. Huth, Commun. 178, 852-867 (2008). M.Kagan,M.D.Schwartz,K.Black,B.Tweedie,JHEP [63] The Atlas Collaboration, Phys. Rev. Lett. 106, 131802 1104, 069 (2011). J. Thaler, K. Van Tilburg, JHEP (2011). The CMS Collaboration, Phys. Lett. B 698 1103, 015 (2011). A. Katz, M. Son, B. Tweedie, Phys. (2011) 196. Rev. D83, 114033 (2011). J. Thaler, Z. Thomas, JHEP