Abstract
We study the NMSSM with universal Susy breaking terms (besides the Higgs sector) at the GUT scale. Within this constrained parameter space, it is not difficult to find a Higgs boson with a mass of about 125 GeV and an enhanced cross-section in the diphoton channel. An additional lighter Higgs boson with reduced couplings and a mass 123 GeV is potentially observable at the LHC. The NMSSM-specific Yukawa couplings and are relatively large and tan is small, such that , , and the top Yukawa coupling are of at the GUT scale. The lightest stop can be as light as 105 GeV, and the fine-tuning is modest. WMAP constraints can be satisfied by a dominantly Higgsino-like LSP with substantial bino, wino, and singlino admixtures and a mass of 60β90 GeV, which would potentially be detectable by XENON100.
1. Introduction
Recently, the ATLAS [1β3] and CMS [4β6] collaborations have presented evidence for a Higgs boson with a mass near 126 GeV (ATLAS) and 125 GeV (CMS in [6]). Interestingly, the best fit to the signal strength in the search channel is about one standard deviation larger than expected in the Standard Model (SM), for ATLAS [1, 2], and for CMS [6].
Since then, several publications have studied the impact of a Higgs boson in the 125 GeV range on the parameter space of supersymmetric (Susy) extensions of the SM [7β38]. Whereas a Higgs boson in the 125 GeV range is possible within the parameter space of the Minimal Susy SM (MSSM) [7β13, 15, 17β21, 25, 29β37], large radiative corrections involving heavy stops are required, which aggravate the βlittle fine-tuning problemβ of the MSSM. In addition, it would be difficult to explain a large enhancement of the diphoton signal strength in the MSSM [15, 34, 37].
Within the Next-to-Minimal Supersymmetric SM (NMSSM [39, 40]), a Higgs boson in the 125βGeV range is much more natural [7, 16, 21, 26, 27, 34]. Additional tree-level contributions and large singlet-doublet mixings in the CP-even Higgs sector can push up the mass of the mostly SM-like Higgs boson and, simultaneously, reduce its coupling to -quarks which results in a substantial enhancement of its branching fraction into two photons [7, 16, 26, 34]. Studies of the parameter space of the general NMSSMβincluding the dark matter relic density and dark matter nucleon cross sectionβwere performed in [34, 38].
An important question is whether these interesting features of the NMSSM survive universality constraints on the soft Susy breaking parameters at the GUT scale. Since this approach imposes severe restrictions on the sparticle masses and couplings, it allows to study whether these would be consistent with present constraints from direct and indirect sparticle searches, the dark matter relic density and dark matter direct detection experiments. Moreover, it allows to make predictions for future searches, both in the sparticle and the Higgs sector.
The (fully constrained) CNMSSM [41, 42] was analysed in [11] with the result that once a relic density in agreement with WMAP [43] is imposed, the Higgs boson mass can barely be above 123βGeV. We find that one should allow for deviations from full universality in the Higgs sector, both for the NMSSM-specific soft Susy breaking terms and the MSSM-like Higgs soft masses like in the MSSM studies in [8, 11, 17, 20]. In analogy to the NUHM version of the MSSM, we will refer to such a model as NUH-NMSSM for nonuniversal Higgs NMSSM. A first study of the NUH-NMSSM was made in [23] which was confined, however, to the more MSSM-like region of the parameter space of the NMSSM involving small values of the NMSSM-specific coupling and hence small NMSSM-specific effects in the CP-even Higgs sector.
In the present paper we study the NUH-NMSSM for large values of the NMSSM-specific coupling (and low ), where the singlet-doublet mixing in the CP-even Higgs sector is large, and we find that the interesting features of the Higgs sector of the NMSSM observed in [7, 16, 21, 26, 27, 34] can remain present, including constraints from searches for squarks and gluinos from ATLAS and CMS [44β46], constraints on the dark matter relic density from WMAP [43] and on the dark matter nucleon cross section from XENON100 [47].
Our results in the Higgs sector originate essentially from the strong mixing between all three CP-even Higgs states in the NMSSM: first, a Higgs boson with a mass in the 125 GeV range can have an enhanced diphoton signal strength up to . Second, a lighter less SM-like Higgs boson exists, with small couplings to electroweak gauge bosons if βGeV (complying with LEP constraints [48]), but a possibly detectable production cross section at the LHC if βGeV. In fact, a strongly enhanced diphoton signal strength of the Higgs boson with its mass in the 125βGeV range is possible only if βGeV. The heaviest CP-even Higgs boson , like the heaviest MSSM-like CP-odd and charged Higgs bosons, has masses in the 250β650βGeV range, while the lightest mostly singlet-like CP-odd Higgs state has a mass in the 160β400βGeV range. These comply with constraints both from -physics and direct Susy Higgs searches also for lower masses due to the low values of considered here and the large singlet component of the lightest CP-odd state.
In the sparticle sector we require masses for the gluino and the first generation squarks to comply with constraints from present direct searches [44β46], but we also study the effect of a reduced sensitivity due to the more complicated decay cascades in the NMSSM [49]. The lightest stop can be as light as ~105βGeV (still satisfying constraints from ATLAS [50] and the Tevatron [51], the latter due to its dominant decay into a chargino and a -quark), and the required fine-tuning among the parameters at the GUT scale remains modest. In the neutralino sector, the mixings among the five states (bino, wino, two Higgsinos, and the singlino) are large. The LSP, with a dominant Higgsino component and a mass of 60β90βGeV, has a relic density complying with the WMAP constraints [43] and a direct detection cross-section possibly within the reach of XENON100 [47]. However, the supersymmetric contribution to the anomalous magnetic moment of the muon is somewhat smaller than desired to account for the deviation of the measurement [52] from the SM.
In the next section we present the analysed parameter space of the NMSSM with boundary conditions at the GUT scale and the imposed phenomenological constraints; our results are given in Section 3 and conclusions in Section 4.
2. The NMSSM with Constraints at the GUT Scale
The NMSSM differs from the MSSM due to the presence of the gauge singlet superfield . In the simplest invariant realisation of the NMSSM, the Higgs mass term in the superpotential of the MSSM is replaced by the coupling of to and and a self-coupling . Hence, in this simplest version the superpotential is scale invariant and given by where hatted letters denote superfields, and the ellipsis denote the MSSM-like Yukawa couplings of and to the quark and lepton superfields. Once the real scalar component of develops a vev , the first term in generates an effective -term,
The soft Susy breaking terms consist of mass terms for the Higgs bosons , , , squarks , , , and sleptons and (where is a generation index): trilinear interactions involving the third generation squarks, sleptons, and the Higgs fields (neglecting the Yukawa couplings of the two first generations): and mass terms for the gauginos (bino), (winos), and (gluinos):
Expressions for the mass matrices of the physical CP-even and CP-odd Higgs statesβafter , , and have assumed vevs , , and and including the dominant radiative correctionsβcan be found in [40] and will not be repeated here. The couplings of the CP-even Higgs states depend on their decompositions into the weak eigenstates , , and , which are denoted by
Then the reduced tree-level couplings (relative to a SM-like Higgs boson) of to quarks, leptons, quarks, and electroweak gauge bosons are
Mixings between the SU(2)-doublet and singlet sectors are always proportional to .
As compared to two independent parameters in the Higgs sector of the MSSM at tree level (often chosen as and ), the Higgs sector of the NMSSM is described by the six parameters,
Then the soft Susy breaking mass terms for the Higgs bosons , , and are determined implicitly by , , and .
In constrained versions of the NMSSM (as in the constrained MSSM) one assumes that the soft Susy breaking terms involving gauginos, squarks, or sleptons are identical at the GUT scale:
In the NUH-NMSSM considered here one allows the Higgs sector to play a special role: the Higgs soft mass terms , , and are allowed to differ from (and determined implicitly as noted above), and the trilinear couplings , can differ from . Hence, the complete parameter space is characterized by where the latter five parameters are taken at the GUT scale.
Subsequently we are interested in regions of the parameter space implying large doublet-singlet mixing in the Higgs sector, that is, large values of (and ) and low values of , which lead naturally to a SM-like Higgs boson in the 125βGeV range [7, 16, 21, 26, 27, 34]. Requiring and , we find with many points for . It is intriguing that with these choices at the weak scale, one obtains for the running couplings at the GUT scale; hence, all 3 Yukawa couplings are close to (but still below) a Landau singularity.
We assume . at the weak scale is required for positive CP-odd Higgs masses squared. We found that, as a consequence, the coupled RG equations imply essentially at the GUT scale. Constraints on the soft Susy breaking parameters depend strongly on the sparticle decay cascades. Using the absence of signal at the LHC in the jets and missing transverse momentum search channels, bounds in the , plane have been derived in the CMSSM with [44β46]. In the NUH-NMSSM, however, we find lighter stops (due to the lower values of implying a larger value of the top Yukawa coupling, which affects the RGE running of the soft Susy breaking stop masses) and a modified neutralino sector which reduces the sensitivity in these search channels [49].
Hence, to start with, we impose only constraints from sparticle searches at LEP [53] and the Tevatron [54, 55], and from stop searches at the Tevatron [51] and the LHC [50]. These imply
In addition we require that the fine-tuning defined in (3.4) satisfies , which implies upper bounds which will be discussed below. However, it is possible that the stronger CMSSM-like constraints in the plane [46] also apply to the NUH-NMSSM considered here. These stronger constraints further reduce the allowed points in the parameter space approximately to but we implemented the constraints from [46] in the , plane exactly. These constraints are used for the points shown in the Figures 1β3 below, but the difference between the constraints (2.12) and (2.13) has practically no impact on our results in the Higgs sector. Remarkably, regardless of the constraints in the , plane, the lightest stop mass can be as low as ~105βGeV.
(a)
(b)
(c)
(d)
(a)
(b)
(a)
(b)
(c)
(d)
Together with these bounds on , , the above constraints on the Higgs sector and the LEP bound on the chargino mass lead to
Appropriate combinations of the bino/wino/higgsino/singlino components of are required in order to satisfy the WMAP and XENON100 bounds. This leads to constraints on βrelevant for the higgsino componentsβas function of , which is proportional to the bino/wino mass terms.
We have scanned the parameter space of the NUH-NMSSM given in (2.10) using a Markov Chain Monte Carlo (MCMC) technique, which yields a very large number of points (~106) satisfying all the phenomenological constraints described below. To do so, we have modified the code NMSPEC [56] inside NMSSMTools [57, 58] in order to allow for and to be used as input parameters at the weak scale and to compute the Higgs soft masses , at the GUT scale; a corresponding version 3.2.0 will be made public soon. In NMSPEC, the two-loop renormalization group equations (RGEs) between the weak and GUT scales are integrated numerically for all parameters. In the presence of boundary conditions both at the weak and the GUT scales as it is the case here, these can be satisfied only through an iterative process. This iterative process is not guaranteed to converge, notably for large Yukawa couplings as in (2.11). In fact, the RGE integration algorithm within the latest public version 3.1.0 of NMSSMTools had to be modified in version 3.2.0 to achieve convergence for large Yukawa couplings.
In the Higgs sector we have used two-loop radiative corrections from [59], and for the top quark pole mass we use βGeV. Our results in the next section use the reduced Higgs production rates (normalized with respect to the SM production rates) in various channels. For gluon-gluon fusion we use the reduced Higgs-gluon coupling as computed in NMSSMTools, which takes care of all colored (s)particles in the loop. For the low values of considered here, the top quark loop dominates by far and leads essentially to as given in (2.7). Since a single particle loop dominates, radiative corrections not considered in NMSSMTools tend to cancel in the ratio to the SM. Likewise, Higgs production rates via associate production with () or vector boson fusion (VBF) are simply proportional to the SM rates rescaled by defined in (2.7). The Higgs branching fractions are computed in NMSSMTools to the same accuracy both for the NMSSM and a SM-like Higgs boson, such that radiative corrections not considered in NMSSMTools tend again to cancel in the ratio to the SM.
Next we turn to the imposed phenomenological constraints. In the Higgs sector we impose constraints from LEP [48], which still allow for a Higgs mass below 114βGeV if its coupling to the boson is reduced. Constraints on Higgs bosons from the LHC are those implemented in version 3.1.0 of NMSSMTools, which are based on public ATLAS and CMS results available at the end of 2011, including constraints from CMS on heavy MSSM-like Higgs bosons decaying to tau pairs [60]. In version 3.2.0, however, we have updated the important search channel with the results from ATLAS [2] and CMS [5]. In order to fit the evidence of both experiments in the channel, we impose , which is satisfied exclusively by for larger values of , and we require a good visibility of in the channel, that is, in both the gluon fusion and VBF production modes.
Also the constraints from -physics are those implemented in version 3.1.0 of NMSSMTools. In spite of charged (resp., CP-odd) Higgs masses as low as ~250 (resp., 160) GeV, these are easily satisfied for low values of or a large singlet component for the lightest CP-odd state (i.e., the couplings of these Higgs bosons to -quarks are hardly enhanced with respect to the SM Higgs).
The dark matter relic density and direct detection cross section of the LSP (the lightest neutralino) are computed with the help of MicrOmegas [61β63] implemented in NMSSMTools. The default constraints are slightly weaker than the most recent ones from WMAP [43], but this has no impact on the viable regions in parameter space (only on the number of points retained). We also apply the bounds from XENON100 [47] on the spin-independent -nucleon cross section (roughly βpb for β 90βGeV).
Since we hardly find light sleptons of the second generation (and again due to the low values of ), the Susy contribution to the anomalous magnetic moment of the muon is below , violating the constraint implemented in NMSSMTools. However, it still improves the discrepancy between the SM and the measured value [52] and can reduce the discrepancy to two standard deviations depending on the employed SM value.
3. Results
Some remarks on our results have already been made above. Notably we have no difficulties to find points in the parameter space satisfying the above constraints, including the dark matter relic density and the direct detection cross section, βGeV and , where we define
The mechanism behind this enhancement has been discussed earlier in [7, 16, 64]: the is strongly enhanced due to a reduced total width (dominated by ) for a small reduced coupling in (2.7), that is, a small value of the mixing angle , in spite of a milder reduction of the partial width . This occurs for large singlet-doublet mixing (which also leads to an increase of ) and requires that the third eigenstate is not decoupled, that is, not too heavy. The enhancement of the also overcompensates a milder reduction of the production cross section of due to singlet-doublet mixing.
The reduction of the total width leads also to a potential increase of the reduced signal rate in the channels (via gluon fusion), in spite of the reduction of the partial widths due to singlet-doublet mixing. A fourth interesting reduced signal cross section is the channel via VBF: which tends to be reduced, however, for a small mixing angle .
Singlet-doublet mixing angles are typically large, if the eigenstates are close in mass. Hence, we should expect that is the larger, the closer is to , that is, the heavier is . Subsequently we consider separately (where is produced via gluon fusion) and (VBF) (where is produced via VBF). In Figure 1 we show , (VBF), , and (VBF) as a function of for a representative sample of ~2000 points in the scanned parameter space of the semiconstrained NMSSM described above. All points satisfy the WMAP bound on the dark matter relic density, the XENON100 bound on , and the stronger lower bound on given in (2.13). Relaxing this bound to the one given in (2.12) does not lead to additional regions in Figure 1.
We see that, as expected, the ratios , (VBF), and can increase with , and can become as large as 2.8 for βGeV. The inverse conclusion does not hold: GeV does not imply . (VBF) is below ~1, and the very small values of (VBF) correspond to the highest values of . The jump in (VBF) near βGeV is caused by a combination of LEP, -physics, and WMAP constraints as well as the condition . For in the range 124β127βGeV, none of these reduced signal rates show a significant dependency on ; corresponding plots would only transcribe the LHC constraints on each rate as a function , but they would not provide additional informations and are hence omitted.
In Figure 1(c) we see that may be below 1. In fact, the βbest fitβ to by ATLAS [1] and by CMS [4] is below 1, whereas the βbest fitβ to is below 1 by CMS [4], but above 1 by ATLAS [1]. We recall that we have . Hence, it may be interesting to see to which extend and ((VBF)) are correlated. This correlation is shown in Figure 2(a), and we see that is possible for up to ~1.5.
Recently, excesses compatible with a Higgs boson in the 125βGeV range have also been observed at the Tevatron [65]. Here, the dominant excess originates from associated production with . The corresponding reduced signal rate for the candidate , () (which is equal to (VBF)) cannot be very small given the observations at the Tevatron. In Figure 2(b) we show () against . We see that is possible only for , but still allows for .
Obviously the reduced signal rates of are also very important. For instance, could be compatible with the excess of events observed by CMS for βGeV in the channel [4]. On the other hand, for β100βGeV the upper bounds from LEP on its reduced coupling to the boson are particularly weak, and a mostly (but not completely) singlet-like could explain the mild excess of events observed there [48, 66, 67]. The corresponding reduced signal cross-sections as a function of are shown in Figure 3. As explained in [16], the reduced signal cross-section in the channel at LEP coincides with .
We see that the reduced signal cross-sections are mostly small for βGeV where the singlet component of is large, but can be as large as ~0.25 for β100βGeV which is interesting given the mild excess of events observed at LEP. On the other hand, for βGeV, , and can become as large as ~0.5 and as large as ~0.9, hence, is potentially detectable.
The Higgs sector of the NMSSM contains a third CP-even state , two CP-odd states and , and, as in the MSSM, a charged Higgs boson . We find that the lightest CP-odd state is mostly singlet-like with a mass in the range 160β400βGeV, and hardly visible at the LHC due to its small production cross-sections. The states , , and all have similar masses in the 250β650βGeV range and would also be difficult to see at the LHC due to the small value of in the region of the parameter space of interest (2.11).
The masses of the sparticles are essentially determined by , , , and . In Figure 4 we show the mass of the lightest first generation squark ( for our choice of parameters) as well as the mass of the lightest (mostly right-handed) stop as a function of the gluino mass . Here it makes a difference whether we impose the weaker bounds (2.12) or the stronger bounds (2.13): points in red satisfy only the weaker bounds while points in green satisfy both constraints. For βGeV, a stop mass as small as 105βGeV is not excluded by present searches at the LHC [50], but could become observable in the near future.
(a)
(b)
It is known that the stop mass has an impact on the fine-tuning with respect to the fundamental parameters of Susy extensions of the SM, due to its impact on the running soft Susy breaking Higgs mass terms. In addition, both are affected by the gluino mass. We have estimated the quantitative amount of fine-tuning with respect to the parameters at the GUT scale following the procedure outlined in [68]. There, a fine-tuning measure was used, where are all parameters at the GUT scale (Yukawa couplings and soft Susy breaking terms). Note that sometimes instead of is used in the definition of , leading to an obvious factor of 2. We find that , shown as a function of and in Figure 5, is dominated as usual by .
(a)
(b)
We see that the fine-tuning can be as low as (with the definition in (3.4)) in the range of smaller stop and gluino masses allowed by the weaker lower bounds (2.12), and still as low as in the range of stop and gluino masses allowed by the stronger bounds (2.13). Both values are an order of magnitude better than in the MSSM [69].
Turning to the neutralino sector we observe, as in the CP-even Higgs sector, large mixing angles involving all 5 neutralinos of the NMSSM. The lightest eigenstate (the LSP) is mostly higgsino-like, but with sizeable bino, wino, and singlino components and a mass in the 60β90βGeV range. The lower bound on follows from the conditions that invisible decays do not spoil the condition , together with the lower LEP bound on chargino masses. Its direct detection cross-section is reduced with respect to pure higgsino-like neutralinos, and can well comply with the constraints from XENON100. In Figure 6 we show the spin-independent neutralino-proton scattering cross-section as a function of . We see that the stronger bounds (2.13) imply βGeV, similar to the range within the general NMSSM obtained in [34]. In particular, the plateau observed in Figure 6 for βGeV and βpb, corresponding to small values of and a mostly bino-like , is excluded by both the XENON100 and the strong LHC Susy constraints. The spin-independent neutralino-proton scattering cross-section can vary over a wide range both above and below the XENON100 limit [47] (which remains to be confirmed by other experiments), but plenty of points would satisfy this constraint and become observable in the future.
It may be interesting to know some of the properties of the Higgs and sparticle sectors beyond the ones shown in the scatter plots above; to this end we show two benchmark points in Table 1. The point (1) has βGeV, the point (2) βGeV, and they differ in the values for and . The branching fractions of are similar for both points: , , .
4. Conclusions
It has already been noted in [7, 16, 21, 26, 27, 34] that the NMSSM can naturally accomodate Higgs bosons in the 124β127βGeV mass range. In addition, the NMSSM can explain excesses in the channel, as well as potential excesses at different values of the Higgs mass (due to the extended Higgs sector). In the present paper we have shown that these features persist in the constrained NMSSM with nonuniversal Higgs sector, designated here as NUH-NMSSM. The dominant deviation from full universality of all soft Susy breaking terms at the GUT scale originates from the need to have .
The following properties of the Higgs sector are peculiar:(i)the signal rate in the channel can be 2.8 as large as the one of a SM-like Higgs boson, provided the mass of the lighter CP-even state is in the 115β123βGeV range;(ii)requiring a visible signal rate in the channel of 0.9 times the SM value allows for a signal rate in the channel about 1.6 as large as the one of a SM-like Higgs boson;(iii)the lighter CP-even state could explain a mild excess of events around 95β100βGeV observed at LEP, or a second visible Higgs boson below ~123βGeV.
In the sparticle sector, the assumption of universality at the GUT scale leads to the following features:(i)the mass of the lightest stop can be as small as 105 GeV, complying with present constraints for βGeV;(ii)the fine-tuning with respect to parameters at the GUT scale remains modest, an order of magnitude below the one required in the MSSM;(iii)the eigenstates in the neutralino sector are strongly mixed, and the lightest neutralino can have a relic density in agreement with WMAP constraints. Its direct detection cross-section can be above or below present XENON100 bounds; most of the points below these bounds should be observable in the near future.
Given the large values of the NMSSM-specific coupling , all scenarios presented here differ strongly from the MSSM (also by the low value of ). The fact that all 3 Yukawa couplings , , and are of at the GUT scale may hint at some strong dynamics present at that scale. It is possible that the deviation from full universality of soft Susy breaking terms at the GUT scale remains confined to ; such possibilities require further studies.
Of course, first of all the present evidence for a Higgs boson in the 124β127βGeV mass range should be confirmed by more data; then possible evidence for non-SM properties of the Higgs sector like an enhanced cross-section in the diphoton channel will show whether the scenarios presented here are realistic.
Acknowledgments
U. Ellwanger acknowledges support from the French ANR LFV-CPV-LHC. C. Hugonie acknowledges support from the French ANRJ TPADMS.