The Higgs sector of the phenomenological MSSM in the light of the Higgs boson discovery

The long awaited discovery of a new light scalar at the LHC opens up a new era of studies of the Higgs sector in the SM and its extensions. In this paper we discuss the consequences of the observation of a light Higgs boson with the mass and rates reported by the ATLAS and CMS collaborations on the parameter space of the phenomenological MSSM, including also the so far unsuccessful LHC searches for the heavier Higgs bosons and supersymmetric particle partners in missing transverse momentum as well as the constraints from B physics and dark matter. We explore the various regimes of the MSSM Higgs sector depending on the parameters MA and tan beta and show that only two of them are still allowed by all present experimental constraints: the decoupling regime where there is only one light and standard--like Higgs boson and the supersymmetric regime in which there are light supersymmetric particle partners affecting the decay properties of the Higgs boson, in particular its di-photon and invisible decays.

1 Introduction and the implications on the pMSSM parameters in the light of the LHC Higgs discovery and constraints. Section 4 has a short conclusion. 2 The theoretical set-up 2

.1 The pMSSM Higgs sector
In the MSSM the Higgs sector is extended to contain five Higgs particles 1 . The lightest h boson has in general the properties of the Standard Model (SM) Higgs boson 2 and is expected to have a mass M h < ∼ 115-135 GeV depending on the MSSM parameters, in particular, the ratio tan β of the vacuum expectation values of the two Higgs doublet fields that break the electro-weak symmetry in the MSSM.
By virtue of supersymmetry, only two parameters are needed to describe the Higgs sector at tree-level. These can be conveniently chosen to be the pseudoscalar boson mass M A and the ratio of vacuum expectation values of the two Higgs fields that break the symmetry, tan β = v 2 /v 1 . However, accounting for the radiative corrections to the Higgs sector, known to play an extremely important role [10], all soft SUSY-breaking parameters which are of O(100) in addition to those of the SM, become relevant. This makes any phenomenological analysis in the most general MSSM a very complicated task. A phenomenologically more viable MSSM framework, the pMSSM, is defined by adopting the following assumptions: i) all soft SUSYbreaking parameters are real and there is no new source of CP-violation; ii) the matrices for the sfermion masses and for the trilinear couplings are all diagonal, implying no flavor change at tree-level; and ii) the soft SUSY-breaking masses and trilinear couplings of the first and second sfermion generations are the same at the electro-weak symmetry breaking scale. Making these three assumptions will lead to only 22 input parameters in the pMSSM: -tan β: the ratio of the vevs is expected to lie in the range 1 < ∼ tan β < ∼ m t /m b ; -M A : the pseudoscalar Higgs mass that ranges from M Z to the SUSY-breaking scale; -µ: the Higgs-higgsino (supersymmetric) mass parameter (with both signs); -M 1 , M 2 , M 3 : the bino, wino and gluino mass parameters; -mq, mũ R , md R , ml, mẽ R : the first/second generation sfermion mass parameters; -A u , A d , A e : the first/second generation trilinear couplings; -mQ, mt R , mb R , mL, mτ R : the third generation sfermion mass parameters; -A t , A b , A τ : the third generation trilinear couplings.
Such a model has more predictability and it offers an adequate framework for phenomenological studies. In general, only a small subset of the parameters appears when looking at a given sector of the pMSSM, such as the Higgs sector in this case. Some of these parameters will enter the radiative corrections to the Higgs boson masses and couplings. At the one-loop level, the h boson mass receives corrections that grow as the fourth power of the top quark mass m t (we use the running MS mass to re-sum some higher order corrections) and logarithmically with the SUSY-breaking scale or common squark mass M S ; the trilinear coupling in the stop sector A t plays also an important role. The leading part of these corrections reads [12] = 3m 4 We have defined the SUSY-breaking scale M S to be the geometric average of the two stop masses (that we take < ∼ 3 TeV not to introduce excessive fine-tuning) and introduced the mixing parameter X t in the stop sector (that we assume < ∼ 3M S ), The radiative corrections have a much larger impact and maximise the h boson mass in the so-called "maximal mixing" scenario, where the trilinear stop coupling in the DR scheme is maximal mixing scenario : In turn, the radiative corrections are much smaller for small values of X t , i.e. in the no mixing scenario : X t = 0.
An intermediate scenario is when X t is of the same order as M S which is sometimes called the typical mixing scenario : These mixing scenarios have been very often used as benchmarks for the analysis of MSSM Higgs phenomenology [13]. The maximal mixing scenario has been particularly privileged since it gives a reasonable estimate of the upper bound on the h boson mass, M max h . We will discuss these scenarios but, compared to the work of Ref. [13], we choose here to vary the scale M S . Together with the requirements on X t in eqs. (4)(5)(6), we adopt the following values for the parameters entering the pMSSM Higgs sector, and vary the basic inputs tan β and M A . For the values tan β = 60 and M A = M S = 3 TeV and a top quark pole of mass of m t = 173 GeV, we would obtain a maximal Higgs mass value M max h ≈ 135 GeV for maximal mixing once the full set of known radiative corrections up to two loops is implemented [14]. In the no-mixing and typical mixing scenarios, one obtains much smaller values, M max h ≈ 120 GeV and M max h ≈ 125 GeV, respectively. Scanning over the soft SUSY-breaking parameters, one may increase these M max h values by up to a few GeV. It is important to note that the dominant two-loop corrections have been calculated in the DR scheme [15] and implemented in the codes Suspect [16] and SOFTSUSY [17] that we will use here for the MSSM spectrum, but also in the on-shell scheme [18] as implemented in FeynHiggs [19]. In general, the results for M h in the two scheme differ by at most 2 GeV, which we take as a measure of the missing higher order effects. Quite recently, the dominant three-loop contribution to M h has been calculated and found to be below 1 GeV [20]. Thus, the mass of the lightest h boson can be predicted with an accuracy of ∆M h ∼ 3 GeV and this is the theoretical uncertainty on M h that we assume.

The various regimes of the pMSSM
The spectrum in the various regimes of the pMSSM Higgs sector [9], depends on the values of M A and also on tan β, and that we will confront to the latest LHC and Tevatron data in this paper.
We start from the decoupling regime [21] that has been already mentioned and which in principle occurs for large values of M A but is reached in practice at M A > ∼ 300 GeV for low tan β values and already at M A > ∼ M max h for tan β > ∼ 10. In this case, the CP-even h boson reaches its maximal mass value M max h and its couplings to fermions and gauge bosons (as well as its self-coupling) become SM-like. The heavier H boson has approximately the same mass as the A boson and its interactions are similar, i.e. its couplings to gauge bosons almost vanish and the couplings to bottom (top) quarks and τ leptons fermions are (inversely) proportional to tan β. Hence, one will have a SM-like Higgs boson h ≡ H SM and two pseudo-scalar (like) Higgs particles, Φ = H, A. The H ± boson is also degenerate in mass with the A boson and the intensity of its couplings to fermions is similar. Hence, in the decoupling limit, the heavier H/A/H ± bosons almost decouple and the MSSM Higgs sector reduces effectively to the SM Higgs sector, but with a light h boson.
The anti-decoupling regime [22] The mass degeneracy is more effective when tan β is large. Here, both the h and H bosons have still enhanced couplings to b-quarks and τ leptons and suppressed couplings to gauge bosons and top quarks, as is the pseudo-scalar A. Hence, one approximately has three pseudo-scalar like Higgs particles, Φ ≡ h, H, A with mass differences of the order of 10-20 GeV.
The intermediate-coupling regime occurs for low values of tan β, tan β < ∼ 5-10, and a not too heavy pseudo-scalar Higgs boson, M A < ∼ 300-500 GeV [9]. Hence, we are not yet in the decoupling regime and both CP-even Higgs bosons have non-zero couplings to gauge bosons and their couplings to down-type (up-type) fermions (as is the case for the pseudoscalar A boson) are not strongly enhanced (suppressed) since tan β is not too large. This scenario is already challenged by LEP2 data which call for moderately large values of tan β.
The vanishing-coupling regime occurs for relatively large values of tan β and intermediate to large M A values, as well as for specific values of the other MSSM parameters. The latter parameters, when entering the radiative corrections, could lead to a strong suppression of the couplings of one of the CP-even Higgs bosons to fermions or gauge bosons, as a result of the cancellation between tree-level terms and radiative corrections [24]. An example of such a situation is the small α eff scenario which has been used as a benchmark [13] and in which the Higgs to bb coupling is strongly suppressed.
Within the plane [M A , tan β], the parameter space in which the above regimes of the pMSSM In addition, we have to consider the SUSY regime, in which some SUSY particles such as the charginos, neutralinos as well as the third generation sleptons and squarks, could be light enough to significantly affect the phenomenology of the MSSM Higgs bosons. For instance, light sparticles could substantially contribute to the loop induced production and decays modes of the lighter h boson [25,26] and could even appear (in the case of the lightest neutralino) in its decay product as will be discussed below.

Higgs decays and production in the pMSSM
For the relatively large values of tan β presently probed at the LHC, tan β > ∼ 7 as discussed below, the couplings of the non-SM like Higgs bosons to b quarks and τ leptons are so strongly enhanced and those to top quarks and gauge bosons suppressed, that the pattern becomes as simple as the following (more details can be found in Ref. [9]): -The Φ = A or H/h bosons in the decoupling/anti-decoupling limit decay almost exclusively into bb and τ + τ − pairs, with branching ratios of, respectively, ≈ 90% and ≈ 10%, and all other channels are suppressed to a level where their branching ratios are negligible.
-The H ± particles decay into fermion pairs: mainly H + → tb and H + → τ ν τ final states for H ± masses, respectively, above and below the tb threshold.
-The CP-even h or H boson, depending on whether we are in the decoupling or antidecoupling regime, will have the same decays as the SM Higgs boson. For M h/H ≈ 126 GeV, the main decay mode will be the bb channel with a ∼ 60% probability, followed by the decays into cc, τ + τ − and the loop induced decay into gluons with ∼ 5% branching ratios. The W W * decay reaches the level of 20%, while the rate for ZZ * is a few times 10 −2 . The important loop induced γγ decay mode which leads to clear signals at the LHC have rates of O(10 −3 ).
In the intense-coupling regime, the couplings of both h and H to gauge bosons and up-type fermions are suppressed and those to down-type fermions are enhanced. The branching ratios of the h and H bosons to bb and τ + τ − final states are thus the dominant ones, with values as in the case of the pseudoscalar A boson. In the intermediate-coupling regime, interesting decays of H, A and H ± into gauge and/or Higgs bosons occur, as well as A/H → tt decays, but they are suppressed in general. Finally, for the rare vanishing-coupling regime when the Higgs couplings to b-quarks and eventually τ -leptons accidentally vanish, the outcome is spectacular for the h boson: the W W * mode becomes dominant and followed by h → gg, while the interesting h → γγ and h → ZZ * decay modes are enhanced.
In the case of the SM-like Higgs particle (that we assume now to be the h boson), there are two interesting scenarios which might make its decays rather different. First we have the scenario with the Higgs bosons decaying into supersymmetric particles. Because most sparticles must be heavier than about 100 GeV, there is no SUSY decays of the h boson except for the invisible channel into a pair of the lightest neutralinos, h → χ 0 1 χ 0 1 . This is particularly true when the gaugino mass universality relation M 2 ∼ 2M 1 is relaxed, leading to light χ 0 1 states while the LEP2 bound, m χ ± 1 > ∼ 100 GeV, still holds. In the decoupling limit, the branching ratio of the invisible decay can reach the level of a few 10%. Decays of the heavier A/H/H ± bosons, in particular into charginos, neutralinos, sleptons and top squarks, are in turn possible. However, for tan β > ∼ 10, they are strongly suppressed. The second scenario of interest occurs when SUSY particles contribute to loop-induced Higgs decays. If scalar quarks are relatively light, they can lead to sizable contributions to the decays h → gg and h → γγ. Since scalar quarks have Higgs couplings that are not proportional to their masses, their contributions are damped by loop factors 1/m 2Q and decouple from the vertices contrary to SM quarks. Only when mQ is not too large compared to M h that the contributions are significant [25]. This is particularly true for thet 1 contributions to h → gg, the reasons being that large X t mixing leads to at 1 that is much lighter than all other squarks and that the h coupling to stops involves a component which is proportional to m t X t and, for large X t , it can be strongly enhanced. Sbottom mixing, ∝ m b X b , can also be sizable for large tan β and µ values and can lead to lightb 1 states with strong couplings to the h boson. In h → γγ decay, there are in addition slepton loops, in particularτ states which behave like scalar bottom quarks and have a strong mixing at high µ tan β, can make a large impact on the decay rate. Besides, chargino loops also enter the h → γγ decay mode but their contribution is in general smaller since the Higgs-χχ couplings cannot be strongly enhanced.
For the evaluation of the decay branching ratios of the MSSM Higgs bosons, we use the program HDECAY [29], which incorporates all decay channels including those involving superparticles and the most important sets of higher order corrections and effects.
Coming to Higgs boson production at the LHC, for a SM-like particle H SM there are essentially four mechanisms for single production [11]. These are gg fusion, gg → H SM , vector boson fusion, qq → H SM qq, Higgs-strahlung, qq → H SM V and tt associated Higgs production, pp → ttH SM . The gg → H SM process proceeds mainly through a heavy top quark loop and is by far the dominant production mechanism at the LHC. For a Higgs boson with a mass of ≈ 126 GeV, the cross section is more than one order of magnitude larger than in the other processes. Again for M H SM ≈ 126 GeV, the most efficient detection channels are the clean but rare H → γγ final states, the modes H → ZZ * → 4 ± , H → W W ( * ) → νν with = e, µ and, to a lesser extent, also H SM → τ + τ − . At the LHC and, most importantly, at the Tevatron one is also sensitive to qq → H SM + W/Z → bb + W/Z with W → ν and Z → , νν.
For the MSSM Higgs bosons, the above situation holds for the h(H) state in the (anti-) decoupling regime. Since AV V couplings are absent, the A boson cannot be produced in Higgs-strahlung and vector boson fusion and the rate for pp → ttA is strongly suppressed. For the Φ = A and h(H) states, when we are in the (anti-)decoupling limit, the b quark will play an important role for large tan β values as the Φbb couplings are enhanced. One then has to take into account the b-loop contribution in the gg → Φ processes which becomes the dominant component in the MSSM and consider associated Higgs production with bb final states, pp → bb + Φ which become the dominant channel in the MSSM. The latter process is in fact equivalent to bb → Φ where the b-quarks are taken from the proton in a five active flavor scheme. As the Φ bosons decay mainly into bb and τ + τ − pairs, with the former being swamped by the QCD background, the most efficient detection channel would be pp → Φ → τ + τ − . This process receives contributions from both the gg → Φ and bb → Φ channels.
These processes also dominate the h/H/A production in the intense coupling regime. In fact, in the three regimes above, when all processes leading to τ + τ − final states are added up, the rate is 2 × σ(gg + bb → A) × BR(A → τ + τ − ). In the intermediate coupling regime, these process have very low cross sections as for 3-5 ≤ tan β ≤ 7-10, the Φbb couplings are not enough enhanced and the Φtt ones that control the gg fusion rate are still suppressed.
Finally, for the charged Higgs boson, the dominant channel is the production from top quark decays, t → H + b, for masses not too close to M H ± = m t −m b . This is true in particular at low or large tan β values when the t → H + b branching ratio is significant.
The previous discussion on MSSM Higgs production and detection at the LHC might be significantly altered if scalar quarks, in particulart andb, are light enough. Indeed, the Hgg and hgg vertices in the MSSM are mediated not only by the t/b loops but also by loops involving their partners similarly to the Higgs photonic decays. The gg → h cross section in the decoupling regime can be significantly altered by light stops and a strong mixing X t which enhances the ht 1t1 coupling. The cross section times branching ratio σ(gg → h) × BR(h → γγ) for the lighter h boson at the LHC could be thus different from the SM case, even in the decoupling limit in which the h boson is supposed to be SM-like [25].
Finally, we should note that in the scenario in which the Higgs bosons, and in particular the lightest one h, decay into invisible lightest neutralinos, h → χ 0 1 χ 0 1 , the observation of the final state will be challenging but possible at the LHC with a higher energy and more statistics. This scenario has recently been discussed in detail in Refs. [5,30].

pMSSM scans and software tools
The analysis is based on scans of the multi-parameter MSSM phase space. The input values of the electro-weak parameters, i.e. the top quark pole mass, the MS bottom quark mass, the electro-weak gauge boson masses, electromagnetic and strong coupling constants defined at the scale M Z , are given below with their 1σ allowed ranges [32], The pMSSM parameters are varied in an uncorrelated way in flat scans, within the following ranges: to generate a total of 6 × 10 7 pMSSM points. The scan range is explicitly chosen to include the various mixing scenarios in the Higgs section discussed in section 2.1: the maximal mixing, no-mixing and typical mixing scenarios. Additional 10 7 points are generated in specialised scans used for the studies discussed later in section 3.5. We select the set of points fulfilling constraints from flavour physics and lower energy searches at LEP2 and the Tevatron, as discussed in Ref. [33], to which we refer also for details on the scans. We highlight here the tools most relevant to this study. The SUSY mass spectra are generated with SuSpect [16] and SOFTSUSY 3.2.3 [17]. The superparticle partial decay widths and branching fractions are computed using the program SDECAY 1.3 [34]. The flavour observables and dark matter relic density are calculated with SuperIso Relic v3.2 [35]. The Higgs production cross sections at the LHC are computed using HIGLU 1.2 [36] for the gg → h/H/A process, including the exact contributions of the top and bottom quark loops at NLO-QCD and the squark loops, and the program bb@nnlo for bb → h/H/A at NNLO-QCD. They are interfaced with Suspect for the MSSM spectrum and HDECAY for the Higgs decay branching ratios. The Higgs production cross sections and the branching fractions for decays into bb, γγ, W W and ZZ from HIGLU and HDECAY are compared to those predicted by FeynHiggs. In the SM both the gg → H SM cross section and the branching fractions agree within ∼ 3%. Significant differences are observed in the SUSY case, with HDECAY giving values of the branching fractions to γγ and W W , ZZ which are on average 9% lower and 19% larger than those of FeynHiggs and have an r.m.s. spread of the distribution of the relative difference between the two programs of 18% and 24%, respectively [5].

Constraints
We apply constraints from flavour physics, anomalous muon magnetic moment, dark matter constraints and SUSY searches at LEP and the Tevatron. These have been discussed in details in Ref. [33]. In particular, we consider the decay B s → µ + µ − , which can receive extremely large SUSY contributions at large tan β. An excess of events in this channels has been reported by the CDF-II collaboration at the Tevatron [37] and upper limits by the LHCb [38] and CMS [39] collaborations at LHC. Recently the LHCb collaboration has presented their latest result for the search of this decay based on 1 fb −1 of data. A 95% C.L. upper limit on its branching fraction is set at 4.5 × 10 −9 [38]. After accounting for theoretical uncertainties, estimated at the 11% level [40] the constraint is used in this analysis. For large values of tan β, this decay can be enhanced by several orders of magnitude so that strong constraints on the scalar contributions can be derived [41], and the small M A and large tan β region can be severely constrained. As already remarked in Ref. [33], the constraints obtained are similar and complementary to those from the dark matter direct detection limits of XENON-100 [42] and searches for the A → τ + τ − decay.
Concerning the relic density constraint, we impose the upper limit derived from the WMAP-7 result [43] 10 accounting for theoretical and cosmological uncertainties [44]. The searches conducted by the ATLAS and CMS collaborations on the √ s = 7 TeV data for channels with missing E T [45,46] have already provided a number of constraints relevant to this study. These have excluded a fraction of the pMSSM phase space corresponding to gluinos below ∼ 600 GeV and scalar quarks of the first two generations below ∼ 400 GeV. These constraints are included using the same analysis discussed in Ref. [33], extended to an integrated luminosity of 4.6 fb −1 .
Then, searches for the MSSM Higgs bosons in the channels h/H/A → τ + τ − [47,48] have already excluded a significant fraction of the [M A , tan β] plane at low M A values, M A < ∼ 200 GeV and tan β < ∼ 10, and larger values of tan β for M A > ∼ 200 GeV. These constraint on the pMSSM parameter space are already important. It is supplemented by the search of light charged Higgs bosons in top decays, t → bH + → bτ ν, performed by the ATLAS collaboration [49] which is effective at low M A values, M A < ∼ 140 GeV, corresponding to M H ± < ∼ 160 GeV. Following the Higgs discovery at the LHC, the lightest Higgs boson in our analysis is restricted to have a mass in the range allowed by the results reported by ATLAS and CMS: where the range is centred around the value corresponding to the average of the Higgs mass values reported by ATLAS and CMS, M h 126 GeV, with the lower and upper limits accounting for the parametric uncertainties from the SM inputs given in eq. (9), in particular the top quark mass, and the theoretical uncertainties in the determination of the h boson mass. It is also consistent with the experimental exclusion bounds.
The impact of the Higgs mass value and its decay rates on the parameters of the pMSSM can be estimated by studying the compatibility of the pMSSM points with the first results reported by ATLAS [1] and CMS [2] at the LHC and also by the Tevatron experiments [3]. Starting from our set of 6 × 10 7 pMSSM points which are pre-selected for compatibility with the constraints discussed above, we consider the two decay channels giving the Higgs boson evidence at the LHC, γγ and ZZ and include also the bb and τ τ channels. In the following, we use the notation R XX to indicate the Higgs decay branching fraction to the final state XX, BR(h → XX), normalised to its SM value. We also compute the ratios of the product of production cross sections times branching ratios for the pMSSM points to the SM values, denoted by µ XX for a given h → XX final state, µ XX = σ(h)×BR(h→XX) σ(H SM )×BR(H SM →XX) . These are compared to the experimental values. For the γγ, and ZZ channels we take a weighted average of the results just reported by the experiments, as given in Table 1  While the results are compatible with the SM expectations within the present accuracy, they highlight a possible enhancement in the observed rates for the γγ channel, where ATLAS and CMS obtain µ γγ = 1.9±0.5 and 1.56±0.43, respectively. In the following, we do not take into account the theoretical uncertainties in the production cross section, which are estimated significant for the main production channel, gg → h [56, 57].  In Figure 3, we show the same [M A , tan β] plane but for different SUSY-breaking scales, M S = 1, 2 and 3 TeV and for the zero, typical and maximal mixing scenarios defined in eqs. (4)(5)(6). As can be seen, the situation changes dramatically depending on the chosen scenario. Still, in the maximal mixing scenario with M S = 3 TeV the size of the M h band is reduced from above, as in this case, already values tan β > ∼ 5 leads to a too heavy h boson, M h > ∼ 129 GeV. In turn, for M S = 1 TeV, the entire space left by the LEP2 and CMS Higgs constraints is covered with many points at tan β > ∼ 20 excluded by the flavor constraint. Nevertheless, the possibility with M S ≈ 1 TeV will start to be challenged by the search for squarks at the LHC when 30 fb −1 of data will be collected by the experiments. In the no-mixing scenario, it is extremely hard to obtain a Higgs mass of M h ≥ 123 GeV and all parameters need to be maximised: M S = 3 TeV and tan β > ∼ 20; a small triangle is thus left over, the top of which is challenged by the flavor constraints. The typical mixing scenario resembles to the no-mixing scenario, with the notable difference that for M S = 3 TeV, the entire space not excluded by the LEP2 and CMS constraints allow for an acceptable value of M h .

The decoupling regime
In the discussion so far, we have adopted the value m t = (173±1) GeV for the top quark mass as measured by the CDF and D0 experiments at the Tevatron [58]. This implicitly assumes that this mass corresponds to the top quark pole mass, i.e. the mass in the on-shell scheme, which serves as input in the calculation of the radiative corrections in the pMSSM Higgs sector and, in particular, to the mass M h . However, the mass measured at the Tevatron is not theoretically well defined and it is not proved that it corresponds indeed to the pole mass as discussed in [59]. For an unambiguous and well-defined determination of the top quark mass, it is appears to be safer to use the value obtained from the determination of the top quark pair production cross section measured at the Tevatron, by comparing the measured value with the theoretical prediction at higher orders. This determination has been recently performed yielding the value of (173.3±2.8) GeV [59] for m pole t . The central value is very close to that measured from the event kinematics but its uncertainty is larger as a result of the experimental and theoretical uncertainties that affect the measurement.
It is interesting to assess the impact of a broader mass range for the top quark. We return to our benchmark scenario with M s = 2 TeV and maximal stop mixing and draw the "M h " bands using the top quark mass values of 170 GeV and 176 GeV corresponding to the wider uncertainty interval quoted above. The result is shown in Figure 4. A 1 GeV change in m t input leads to a ∼1 GeV change in the corresponding M h value. The smaller value of m t would open up more parameter space as the region in which M h > ∼ 129 GeV will be significantly reduced. In turn, for m t = 176 GeV, the corresponding h boson mass increases and the dark-blue area quite significantly shrinks, as a result. It must be noted that for m t = 170 GeV, the no-mixing scenario would be totally excluded for M S < ∼ 3 TeV, while in the typical mixing scenario only a small area at high tan β will remain viable. For m t = 176 GeV significant [M A , tan β] regions that was excluded when taking the ± 1 GeV uncertainty for top mass value becomes allowed. The impact of the value of m t is thus extremely significant. This is even more true in constrained scenarios, where the top mass also enters in the evaluation of the soft SUSY-breaking parameters and the minimisation of the scalar potential. To visualise the impact of m t , we have repeated the study presented in Ref. [4], presenting the maximal M h value reached when scanning over all the parameters of the minimal SUGRA, AMSB and GMSB models. Figure 5 shows the result with the M max h value as a function of M S taking m t =173±3 GeV. While for m t = 173 GeV, there is no region of the parameter space of the mAMSB and mGMSB models which satisfies M h > ∼ 123 GeV, for M S < ∼ 3 TeV assumed in [4], and the models are disfavoured, using m t =176 GeV, the regions of these mAMSB and mGMSB models beyond M S = 2 TeV become again viable. This will be also the case of some of the variants and even more constrained mSUGRA scenarios. Further, even for m t =173 GeV, if we move the M S upper limit from the 3 TeV boundary adopted in Ref. [4] to M S = 5 TeV, these models have region of their parameters compatible with the LHC Higgs mass.
Finally, we comment on the impact of increasing the M h allowed range from 123 GeV ≤ M h ≤ 127 GeV as was done in Ref. [4] relying on the 2011 LHC data, to the one adopted here, 123 GeV ≤ M h ≤ 129 GeV, in the various constrained models discussed in that reference (and to which we refer for the definition of the models and for the ranges of input parameters that  This limit can be strengthened by the same τ + τ − search performed by the ATLAS collaboration [47] and also by the t → H + b search in top decays [49] which is effective for GeV and which, as can be seen in Fig. 7, excludes large tan β values for which BR(t → bH + ) is significant. Put together, these constraints exclude entirely both the anti-decoupling and intense coupling regimes. Would remain then, the intermediate coupling regime with tan β ≈ 5-8 when the LEP2 constraint is also imposed. Depending on the mixing scenario, most of it will be excluded by the M h ≈ 126 GeV constraint (see Fig. 3). actually the H state, while the lighter h boson has suppressed couplings to W/Z bosons and top quarks, allowing it to escape detection. In this case, the H couplings to bottom quarks should not be too enhanced, tan β < ∼ 8, not to be in conflict with the τ + τ − and t → bH + searches above. For the H boson to be SM-like, one should have M A ≈ M H ≈ 126 GeV and not too low tan β values, tan β > ∼ 7-10. One is then in the borderline between the antidecoupling and the intermediate coupling regimes. We have searched for points in which indeed M H ≈ 126 GeV with couplings to V V states, g HV V > ∼ 0.9, such that the H → ZZ and H → γγ (which mainly occurs through a W -boson loop) decays are not suppressed compared to the measured values by ATLAS and CMS given in Table 1. In our scan, out of the 10 6 points, before imposing any LHC-Higgs constraint, only ≈ 20 points fulfilled the above requirements. These points are then completely excluded once the flavor constraints, in particular those from the b → sγ radiative decay, are imposed. Hence, the possibility that the observed Higgs particle at the LHC is not the lightest h particle appears highly unlikely according to the result of our scan of the parameter space. Combining the h/A → τ + τ − and the t → bH + constraints and including the results on the new 8 TeV data should further constrain the parameter space and completely exclude this scenario.
Finally, the vanishing coupling regime is strongly disfavoured by the LHC and Tevatron data that are summarised in Table 1. The observation of H → ZZ final states by the ATLAS and CMS collaborations rules out the possibility of vanishing hV V couplings. The reported excess of events in the qq → V H → V bb process by the CDF and D0 collaborations seem also to rule out both the vanishing hbb and hV V coupling scenarios. However, there is still the possibility that these couplings are smaller than those predicted in the SM case, in particular because of the effects of SUSY particles at high tan β [24]. We are then in the SUSY-regime to which we turn now.

The SUSY regime
In the SUSY regime, both the Higgs production cross section in gluon-gluon fusion and the Higgs decay rates can be affected by the contributions of SUSY particles. This makes a detailed study of the pMSSM parameter space in relation to the first results reported by the ATLAS and CMS collaborations especially interesting for its sensitivity to specific regions of the pMSSM parameter space. In particular, the branching fraction for the γγ decay of the h state is modified by Higgs mixing effects outside the decoupling regime as was discussed above, by a change of the hbb coupling due to SUSY loops [24], by light superparticle contributions to the hγγ vertex [6,25,26] and by invisible h decays into light neutralinos [28].
We study these effects on the points of our pMSSM scan imposing the LHC results as constraints. The numerical values adopted in the analysis are given in Table 1, assuming in the following on that the observed particle is the h state. First, we briefly summarise the impact of the SUSY particles on the Higgs decay branching fractions, staring from invisible decays, and production cross sections. Then we discuss our finding on the impact of the LHC and Tevatron data on the pMSSM parameters.

Invisible Higgs decays
Despite the fact that the discovered particle has a sufficient event rate in visible channels to achieve its observation, it is interesting to consider the regions of parameter space in which invisible Higgs decays occur. This scenario has recently been re-considered in [5,30]. Besides the value of M h , the invisible branching ratio BR(h → χ 0 1 χ 0 1 ) is controlled by four parameters: the gaugino masses M 1 and M 2 , the higgsino parameter µ and tan β. They enter the 4 × 4 matrix Z which diagonalises the neutralino mass matrix. They also enter the Higgs coupling to neutralinos which, in the case of the LSP, is if we assume the decoupling limit not to enhance the h → bb channel which would significantly reduce the invisible decay. In this coupling, Z 11 , Z 12 are the gaugino components and Z 13 , Z 14 the higgsino components. Thus, the coupling vanishes if the LSP is a pure gaugino, |µ| M 1 leading to m χ 0 1 ≈ M 1 , or a pure higgsino, M 1 |µ| with m χ 0 1 ≈ |µ|. For the invisible decay to occur, a light LSP, m χ 0 1 ≤ 1 2 M h is required. Since in the pMSSM, the gaugino mass universality M 2 ≈ 2M 1 is relaxed, one can thus have a light neutralino without being in conflict with data. The constraint from the Z invisible decay width measured at LEP restricts the parameter space to points where theχ 0 1 is bino-like, if its mass is below 45 GeV, and thus to relatively large values of the higgsino mass parameter |µ|. Since a large decay width intoχ 0 1χ 0 1 corresponds to small values of |µ|, this remove a large part of the parameter space where the invisible Higgs decay width is sizable. Still, we observe invisible decays for 45 GeV< Mχ0 < M h 0 /2 and |µ| < 150, corresponding to a combination of parameters where theχ 0 1 is a mixed higgsino-gaugino state [5]. These pMSSM points are shown in the [M 1 , µ] plane in the left panel of Fig. 8. If the LSP at such a low mass were to be the dark-matter particle, with the relic density given in eq. (12), it should have an efficient annihilation rate into SM particles. The only possible way for that to occur would be χ 0 1 χ 0 1 annihilation through the s-channel light h pole 3 [60] which implies that m χ 0 1 < ∼ 1 2 M h to still have a non-zero invisible branching ratio, as shown in the right panel of Fig. 8, where the pMSSM points satisfying BR(h → χ 0 1 χ 0 1 ) ≥ 5% are shown in the plane [m χ 0 1 , log 10 (Ωh 2 )]. However, because the partial decay width Γ(h → χ 0 1 χ 0 1 ) is suppressed by a factor β 3 near the M h ≈ 2m χ 0 1 threshold, with the velocity β = (1 − 4m 2 χ 0 1 /M 2 h ) 1/2 , the invisible branching fraction is rather small if the WMAP dark matter constraint is to hold. MSSM light neutralinos compatible with claims of direct detection dark matter signals are also consistent with collider bounds [61].
3.5.2 Sparticle effects on the hbb coupling SUSY particles will contribute to the hbb coupling as there are additional one-loop vertex corrections that modify the tree-level Lagrangian that incorporates them [24]. These corrections involve bottom squarks and gluinos in the loops, but there are also possibly large corrections from stop and chargino loops. Both can be large since they grow as µ tan β or A t µ tan β [24] .
Outside the decoupling limit, the reduced bb couplings of the h state are given in this case by and can be thus significantly reduced or enhanced 4 depending on the sign of µ and, possibly, also A t . This is exemplified in the left panel of Fig. 9, where the ratio R bb ≡ BR(h → bb)/BR(H SM → bb) is shown as a function of the parameter µ tan β before the constraints of Table 1. The two branches in the histogram are due to the sbottom and stop contributions in which R bb is increased or decreased depending on the sign of µ.
A deviation of the partial h → bb width will enter the total Higgs width, which is dominated by the bb channel, and change the R values for the different Higgs decay channels. A reduction of R bb would thus lead to an enhancement of the γγ and the W W /ZZ branching fractions. Figure 10 shows the values of R bb and R γγ in which we observe a highly anti-correlated variation of the two ratios, with the exception of the cases where the opening of the decay h → χχ suppresses both branching fractions. The preliminary results from LHC and the Tevatron are overlayed.

Sparticle contributions to the hgg and hγγ vertices
Scalar top quarks can alter significantly the gg → h cross section as well as the h → γγ decay width [25]. The current eigenstatest L ,t R mix strongly, with a mixing angle ∝ m t X t , so that 3 The other possible channels are strongly suppressed or ruled out. The co-annihilation with charginos, heavier neutralinos and staus is not effective as these particles need to be heavier than ≈ 100 GeV and thus the mass difference with the LSP is too large. The annihilation through the A-pole needs M A ≈ 2m χ 0 1 < ∼ M h and sizable tan β values, which is the anti-decoupling regime that is excluded as discussed above. Remains then the bulk region with staus exchanged in the t-channel in χ 0 1 χ 0 1 → τ + τ − (sbottoms are too heavy) which is difficult to enhance as the LSP is bino-like. 4 These corrections also affect the Higgs production cross sections in the channels gg + bb → Φ. However, in the cross sections times branching ratios for the τ + τ − final states, they almost entirely cancel as they appear in both the production rate and the total Higgs decay width [57].  for large X t = A t − µ/ tan β values 5 there is a lighter mass eigenstatet 1 which can be much lighter than all other scalar quarks, mt 1 M S . The coupling of the h boson to thet 1 states in the decoupling regime reads In the no-mixing scenario X t ≈ 0, the coupling above is ∝ m 2 t and the scalar top contribution to the hgg amplitude is small, being damped by a factor 1/m 2 t 1 and interferes constructively with the top quark contribution to increase the gg → h rate. However, since in the no-mixing 5 On should assume X t values such that A t < ∼ 3M S to avoid dangerous charge and colour breaking minima. In addition, if X t > ∼ √ 6M S , the radiative corrections to the h boson mass become small again and it would be difficult to attain the value M h ≈ 126 GeV. scenario M S = √ mt 1 mt 2 has to be very large for the h boson mass to reach a value M h ≈ 126 GeV, the stop contribution to the hgg vertex, ∝ m 2 t /M 2 S , is very small. In the maximal mixing scenario, X t ≈ √ 6M S , it is the last component of g ht 1t1 which dominates and becomes very large, ∝ −(m t X t /mt 2 ) 2 . However, in this case, the large contribution of a light stop to the hgg amplitude interferes destructively with the top quark contribution and the gg → h cross section is suppressed. For mt 1 ≈ 200 GeV and X t ≈ 1 TeV, we obtain a factor of two smaller gg → h rate. In the case of sbottom squarks, the same situation may occur for large sbottom mixing X b = A b − µ tan β. However, for large value of M S , it is more difficult to obtain a small enough mb 1 state to significantly affect the gg → h cross section. In the case of the hγγ decay amplitude, there is the additional SM contribution of the W boson, which is in fact the dominant. Also, it has the opposite sign to that from the top quark and, hence, when stops are light and have a strong mixing, they will tend to increase the hγγ amplitude. However, because the W contribution is by far the largest, the stop impact will be much more limited compared to the ggh case and we can expect to have only a ≈ 10% increase of the h → γγ decay rate for mt 1 ≈ 200 GeV and X t ≈ 1 TeV [25][26][27]. Therefore, for light and strongly mixed stops, the cross section times branching ratio µ γγ is always smaller unity and relatively light stops do not entail an enhancement of the γγ yield. The sbottom contribution ) (right). We impose R bb > 0.9, to remove the effects due to the changes of the total width through the bb channel.
to the hγγ vertex is also very small, for the same reasons discussed above in the case of the hgg amplitude, and also because of its electric charge, − 1 3 compared to + 2 3 for stops. Other charged particles can also contribute to the h → γγ rate [26]. The charged Higgs bosons have negligible contributions for m H ± > ∼ 200 GeV. Charginos contribute to the hγγ vertex and, because of their spin 1 2 nature, they contribution is only damped by powers of M h /m χ ± . However, the hχ ± 1,2 χ ∓ 1,2 couplings are similar in nature to those of the LSP given in eq. (14) and cannot be strongly enhanced. As a result we expect contributions at most of the order of 10% even for mass values m χ ± 1 ≈ 100 GeV (see Fig. 11). Charged sleptons have in general also little effect on the hγγ vertex, with the exception of staus [6]. These behave like the bottom squarks. At very large µ tan β values, the splitting between the twoτ states becomes significant and their couplings to the h boson large. Sinceτ 1 can have a mass of the order of a few 100 GeV, without affecting the value of M h , its contribution to the hγγ amplitude may be significant for largte values of X τ (see Fig. 11).

Impact of the LHC data
Now, it is interesting to perform a first assessment of the compatibility of the LHC and Tevatron data with the MSSM and analyse the region of parameter favoured by the observed boson mass and rate pattern (see also [5,62]). Despite the preliminary character of the results reported by the LHC collaborations and the limited statistical accuracy of these first results, the study is a template for future analyses. In this analysis, we computing the χ 2 probability on the observable of Table 1 for each accepted pMSSM points. For the bb and τ + τ − channels, in which no evidence has been obtained at the LHC, we add the channel contribution to the total χ 2 only when their respective µ value exceeded 1.5 and the pMSSM point becomes increasingly less consistent to the limits reported by CMS. In order to investigate the sensitivity to the inputs, we also compare the results by including or not the bb, for which a tension exists between the CMS limit and Tevatron results, and the τ + τ − rate. Figure 12 shows the region of the [X t , mt 1 ], [X b , mb 1 ] and [M A , tan β] parameter space where pMSSM points are compatible with the input h boson mass and observed yields. In particular, we observe an almost complete suppression for low values of the sbottom mixing parameter X b . The distributions for some individual parameters which manifest a sensitivity are presented in Figure 13, where each pMSSM point enters with a weight equal to its χ 2 probability. Points having a probability below 0.15 are not included. The probability weighted distributions obtained from this analysis are compared to the normalised frequency distribution for the same observables obtained for accepted points within the allowed mass region 122.5 < M H <127.5 GeV. We observe that some variables are significantly affected by the constraints applied. Not surprisingly, the observable which exhibits the largest effect is the product µ tan β, for which the data favours large positive values, where the γγ branching fraction increases and the bb decreases as discussed above. On the contrary, it appears difficult to reconcile an enhancement of both µ γγ and µ bb , as would be suggested by the central large value of µ bb = 1.97±0.72 recently reported by the Tevatron experiments [3]. Such an enhancement is not observed by the CMS collaboration and the issue is awaiting the first significant evidence of a boson signal in the bb final state at the LHC and the subsequent rate determination. The tan β distribution is also shifted towards larger value as an effect of the Higgs mass and rate values. We also observe a significant suppression of pMSSM points with the pseudo-scalar A boson mass below ∼450 GeV. This is due to the combined effect of the A → τ + τ − direct searches and B s → µ + µ − rate, which constrain the [M A − tan β] plane to low tan β value for light A masses, by the shift to µ tan β from the Higgs rates disfavouring the low tan β region and by the suppression of the non-decoupling regime.
In quantitative terms, we observe that 0.06 (0.50) of the selected pMSSM points are compatible with the constraints given in Table 1 at the 68% (90%) confidence level. If we remove the constraint on the upper limit constraint on the bb and τ + τ − rates, the fraction of points accepted at the 90% C.L. does not change significantly, at 0.56, but that at the 68% C.L. doubles to 0.12. On the contrary, if we replace the CMS upper limit for µ bb with the µ bb result of the Tevatron experiments for M H = 125 GeV [3], the fraction of accepted points at 68% C.L. drops below 0.005. This highlights the tension which will be created in the pMSSM by a simultaneous excess in the γγ and bb channels, excess which cannot be adequately described in the pMSSM, as discussed above (see Figure 10).

Conclusions
The implications of the new boson observation by the ATLAS and CMS collaborations for the phenomenological MSSM have been outlined. The study has been based on broad scans over the pMSSM parameter space where points have been preselected based on constraints from electroweak and flavour physics, dark matter and searches at LEP2 and the LHC. Various scenarios for the stop mixing parameter X t (maximal, typical and zero-mixing) and representative values of the soft SUSY-breaking scale M S (1, 2 and 3 TeV) have been confronted with the Higgs mass range compatible with LHC results, accounting for systematic uncertainties. In order to obtain M h in the mass range 123 GeV≤ M h ≤ 129 GeV, large values of M S and/or X t are required. In particular, the M h constraints are sensitive to the value of the top quark mass for which the value extracted from the top quark pair production cross section has a more unambiguous definition but larger uncertainties.
The various regimes of the pMSSM Higgs sector have been examined in the [M A , tan β] parameter. Of these regimes, only the decoupling regime, where the lighter h boson has almost SM-like properties and the heavier Higgs particles decouple from gauge bosons, and the SUSY regime survives all constraints. The anti-decoupling regime where the H state plays the role of the SM Higgs boson, the intense coupling regime in which there are three light states h, H and A, the vanishing coupling regime in which the h coupling to bottom quarks or gauge bosons are very strongly suppressed, and most of the intermediate coupling regime with relatively low M A and tan β values, are excluded by the present data. In the SUSY regime light superparticles may affect the production and decay rates of the h boson. Light neutralinos may lead to invisible h boson decays, light stop and sbottom quarks affect the hbb couplings and the production cross section in the dominant gluon-gluon fusion mechanism, and light squarks, τ -sleptons and charginos may affect the h → γγ decay mode.
We have confronted these possibilities with the recent LHC results and find that a significant fraction of pMSSM points in our scan compatible with them, including a possible enhancement of the γγ rate. Improved precision in the experimental measurements and sensitivity to the direct searches for the heavier Higgs bosons and supersymmetric particle partners at the LHC will provide the basis for clarifying the relation between the newly discovered scalar sector and physics beyond the Standard Model.