BSM Higgs results from ATLAS and CMS

Searches for Higgs bosons in different extensions of the Standard Model (SM) are presented. These include the Minimal Supersymmetric extension of the SM (MSSM), the next-to-MSSM (NMSSM), models with additional scalar singlets, doublets, or triplets, and generic searches for models with couplings modified with respect to the SM or for non-SM Higgs boson decay channels. Results are based on data collected by the ATLAS and CMS experiments in 2011 and 2012 at the LHC. No excess is found in any of the searches and thus the resulting exclusion limits are given.


Introduction
A large number of searches for beyond-the-SM (BSM) Higgs bosons have been performed by the ATLAS [1] and CMS [2] experiments at the LHC [3], using collision data at a center-of-mass energy of 7 TeV (2011) and 8 TeV (2012). The discovery of a particle with a mass of 125 GeV compatible with a Higgs boson by the AT-LAS and CMS experiments [4,5] leads to new questions -in particular whether it is the SM Higgs boson or one of typically many Higgs bosons predicted by extensions of the Standard Model. Experimentally, this implies measuring the properties of the new boson (not covered here) and searching for additional Higgs bosons. In the following an overview of the most recent results of searches for BSM Higgs bosons using an integrated luminosity of at least 5 fb −1 is presented.

MSSM Higgs boson searches
The MSSM employs a type-II 2-Higgs doublet model (2HDM), which predicts the existence of three neutral Higgs bosons (h, H, A) and a charged pair (H ± ). At treelevel, the Higgs sector of the MSSM can be described by only two additional parameters, e.g. by one of the masses of the heavy Higgs bosons and tan β, the ratio of the vacuum expectation values of the two Higgs doublets. In the MSSM, the recently discovered boson with m ≈ 125 GeV can be identified with the h boson for most of the m A − tan β region not excluded by direct MSSM Higgs boson searches and also with the H particle if the charged Higgs boson mass is below 150 GeV and tan β is small [6]. Since the couplings of the MSSM Higgs bosons to b and t quarks and tau leptons are enhanced for large tan β most a e-mail: martin.flechl@cern.ch searches for these bosons focus on Higgs boson production in association with b and t quarks and decay modes involving b, t and τ particles.

H + → cs
ATLAS searches for charged Higgs bosons produced in semileptonic tt events with t → bH ± and H ± → cs in 4.7 fb −1 of 7 TeV data [7]. The analysis requires an isolated lepton, 4 jets (two of them b-tagged), missing transverse energy (E miss T ) greater than 20 (30) GeV and the transverse mass m T of the lepton and E miss T to satisfy m T > 30 GeV (m T + E miss T > 60 GeV) in the muon (electron) channel. A kinematic fit is applied to the remaining events with the goal to find a second mass peak in the dijet distribution due to H + → cs events. Good agreement between data and SM expectation is observed, leading to a limit on the branching ratio B(t → bH + )=(5 − 1)% for m H ± = 90 − 150 GeV, assuming B(H + → cs)=100%, as shown in figure 1.

H + → τν
Both ATLAS and CMS search for charged Higgs bosons produced in tt events that decay to H + → τν. The ATLAS analysis [8] is based on 4.6 fb −1 of 7 TeV data and studies the final states τ lep +jets, τ had +lepton and τ had +jets where the subscript indicates whether the tau lepton decays to leptons or hadrons+neutrino. The τ had +jets channel is the most sensitive one for most of the H + mass range. The selection starts with a τ + E miss T trigger and requires the presence of exactly one hadronically decaying tau lepton, at least four jets (at least one of them b-tagged), no electrons or muons, E miss T > 65 GeV, a high significance of the missing transverse energy (quantified by its ratio to the square root of the sum of transverse momenta of reconstructed tracks) and a topology consistent with the presence of a hadronic top quark decay. The final discriminant used is a transverse mass built from the visible tau lepton decay products and the missing momentum vector. All backgrounds are estimated in a data-driven way by replacing muons in µ+jets collision data with simulated tau leptons for backgrounds with actual τ leptons, and by deriving data-driven corrections for the misidentification probabilities for backgrounds where jets or electrons are misidentified as τ leptons.
For all final states, the data agree with the SM expectation and an upper limit of B(t → bH + ) = (5 − 1)% assuming B(H + → τν) = 100% is set for m H ± = 90 − 160 GeV. In a different approach, lepton universality in tt events is tested by observing the ratio of event yields with electrons or muons and with and without tau leptons in the final state which would be enhanced by the presence of a charged Higgs boson [9]. This brings down the limit after combining with the τ had +jets final state to about B(t → bH + ) = (3 − 1)%. Interpreting this limit in the context of the m max h of the MSSM [10], a large portion of the parameter space for a light H ± can be excluded, see figure 2.

Φ → bb
In the MSSM, the mode Φ → bb (Φ = h, H, A) is dominant for most regions of the parameter space. The CMS search [12] uses up to 4.8 fb −1 of 7 TeV data and focuses on b-associated production, pp → bΦ → bbb, in two separate channels: The first requires three tight b-tagged jets and is split into two analyses searching for m Φ either below or above 180 GeV. The second employs looser b-tagging requirements but demands that a muon overlaps with the leading b-tagged jet. The final discriminant is the invariant mass of the two leading b-tagged jets. Backgrounds are estimated from data, using control samples with 1 or

Φ → ττ, Φ → µµ
The CMS analysis uses 17 fb −1 of 7 TeV and 8 TeV data to search for MSSM Higgs boson decays to a tau lepton pair [13], and 5 fb −1 of 7 TeV data for decays to a muon pair [14]. For the former, three different combinations of tau lepton decays are considered: τ lep τ had , τ e τ µ and τ µ τ µ . The ATLAS analysis [15] in addition studies   the τ had τ had final state, but does not use τ µ τ µ . For both ATLAS and CMS, each final state is split into two categories with and without additional b-tagged jets aiming to exploit the different production mechanisms (b-associated and gg-fusion). The b-tagged categories provide a higher signal-over-background ratio but a smaller number of expected signal events. The final discriminant in all cases is an estimator, "Missing Mass Calculator" for ATLAS and "SVFit" for CMS, for the ditau mass based on different algorithms for finding the most likely value given the unknowns due to the neutrinos in the decay chain. The main background in all categories are Z → ττ events which are estimated using a technique replacing muons in Z → µµenhanced collision data by simulated tau leptons. Most other relevant backgrounds are normalized in control regions.
No significant excess is observed and exclusion limits are produced. ATLAS provides model-independent cross section limits for both gluon-gluon fusion and b-quark associated production (see figure 4); and both experiments interpret their results in the context of the m max h scenario of the MSSM (the CMS limit is shown in figure 5). CMS excludes m A < 125 GeV (in the region not yet excluded by LEP) as well as values of tan β > 5 for m A < 225 GeV; the exclusion region extends up to m A = 800 GeV and tan β > 50.  H → WW → eνµν decay mode with 0 or 2 additional jets and uses 13 fb −1 of 8 TeV data [16]. The preselection is in common with the SM H → WW search. To maximize sensitivity, an artificial neural net is employed using kinematic information about the event. No evidence for a Higgs boson in the mass range of 135-300 GeV is found. Exclusion limits are set for type-I and type-II 2HDM for a set of fixed values of tan β ranging from 1 to 50 in the m H -cos α plane (an example is shown in figure 6). Assuming that other Higgs bosons or additional particles are too heavy to interfere the tree-level cross section depends only on these three unknown parameters.  [19]. No significant excess is observed, and an additional heavy Higgs boson with SM couplings is excluded up to a mass of m H = 600 GeV; however, a cross section times branching ratio less than 10 − 70% (depending on m H ) of the SM prediction is still allowed, see figure 7.

Higgs mass [GeV]
Recent heavy Higgs boson searches in the H → ZZ channel focus on the final states 2l2q [20], 2l2ν [18] and 4l + 2l2τ [21] for CMS, and 4l [22] for ATLAS. The data agree well with the SM expectation, and the most stringent limits can be set in the 4l + 2l2τ analysis and are shown in figure 8 for m H = (180−500) GeV, a cross section times branching ratio larger than 20% of the SM prediction is excluded.

Invisible Higgs, ZH
The ATLAS search for invisible Higgs boson decays [23] considers Higgs bosons produced in association with a Z boson in 18 fb −1 of 7 TeV and 8 TeV data. To tag the event, an electron or muon pair consistent with the Z boson mass is required. Events with additional leptons or jets are rejected. The missing transverse energy is used as final discriminant. No excess is observed, and assuming m H = 125 GeV and the SM cross section for ZH production, a branching ratio for invisible Higgs boson decays larger than 65% is excluded. In addition, upper limits between 30−7 fb are set on the cross section times branching ratios for invisible Higgs boson decays in the mass range m H = (115 − 300) GeV, see figure 9.

3.3
H → a 0 a 0 → 4γ ATLAS uses 4.9 fb −1 of 7 TeV data to search for Higgs boson decays to very light CP-odd scalars (m a 0 of a few hundred MeV) which in turn decay to photon pairs [24]. This is motivated by composite-Higgs models or the NMSSM but the results are interpreted in a model-independent way. Such a process leads to two highly collimated a 0 → γγ decays, which can mimick a diphoton event. The analysis requires such a diphoton pair using looser requirements for the shower shapes for photon identification than e.g. in the SM H → γγ search due to two overlapping electromagnetic showers for the signal hypothesis. No significant excess is observed and upper limits of a few times 0.1 pb on the cross section times branching ratios is set for the

Doubly-charged Higgs Φ ±±
Both ATLAS and CMS use about 5 fb −1 of 7 TeV data to search for doubly-charged Higgs bosons in decays to same-sign lepton pairs e ± e ± , e ± µ ± , µ ± µ ± . CMS, in addition, also looks for all possible combinations with tau leptons. The CMS analysis [25] investigates both pair production of and associated production with a singlycharged Higgs boson. The data agree with the SM expectation and upper limits on the cross section times branching ratio for m Φ ±± = 130 − 500 GeV are reported for different assumptions on the branching ratio of the doubly-charged Higgs boson. This result is translated in lower bounds on m Φ ±± , see figure 11 for an example. For a branching ratio of 1 for the different lepton pair decays, doubly-charged
The ATLAS analysis [26] searches for a narrow peak in the mass spectrum of same-sign lepton pairs. No excess is observed and the resulting limits on the cross section times branching ratio lead to lower bounds of about 400 GeV on the Φ mass, assuming pair producion and a branching ratio of 1 to the investigated final states.

Higgs in fermiophobic and SM4 models
CMS searches for Higgs bosons in SM-like models with a fourth generation of fermions using up to 10 fb −1 of 7 TeV and 8 TeV data [27]. In the investigated benchmark scenario, the effective Higgs coupling to gluons is enhanced by a factor of 4-9 (depending on m H ) while the decay to two photons is almost entirely suppressed. The search thus focuses on Higgs bosons produced in gg-fusion and decaying to WW, ZZ and ττ. No significant excess compatible with the signal hypothesis is observed and when combining the channels considered, consequently the SM4 benchmark scenario is excluded for Higgs boson masses between 110 GeV and 600 GeV (see figure 12).
In fermiophobic models, the Higgs coupling to fermions is suppressed. The CMS study [27], using 10 fb −1 of 7 TeV and 8 TeV data, thus focuses on Higgs boson decays to photons in 9 different categories. The analysis excludes a fermiophobic Higgs boson with a mass in the range 110-147 GeV.

Conclusions
A large number of BSM Higgs boson searches have been performed by the ATLAS and CMS experiments. No significant excess is observed, and various cross-section limits and exclusion regions for the parameter space of several models have been provided. In particular the MSSM is becoming heavily constrained: A new lower limit of m A > 125 GeV has been set for the m max h scenario (when including the region excluded by LEP), and for 125 < m A [GeV] < 225 only a tan β of about 2-5 is still allowed. The limits extend up to m A = 800 GeV and tan β = 50. Strong cross section limits have been set for the existence of additional Higgs bosons with SM-like properties covering a large range of masses up to 1 TeV. Furthermore, general 2HDMs, the NMSSM as well as type-II seesaw mechanism models have been constrained.
Very few of these analyses have analysed the full Run-I LHC dataset, and in addition predictions of several well-motivated BSM models have not been experimentally tested at the LHC. It is thus still possible that BSM Higgs physics is hiding in the current data -and certainly, additional data collected at √ s = 13 TeV or higher at the LHC beginning from 2015 will greatly enhance the sensitivity to BSM Higgs bosons.