Addendum to "Constraints on and future prospects for Two-Higgs-Doublet Models in light of the LHC Higgs signal"
AAddendum to “Constraints on and future prospects forTwo-Higgs-Doublet Models in light of the LHC Higgs signal”
B´eranger Dumont , ∗ John F. Gunion , † Yun Jiang , ‡ and Sabine Kraml § (1) Laboratoire de Physique Subatomique et de Cosmologie,Universit´e Grenoble-Alpes, CNRS/IN2P3,53 Avenue des Martyrs, F-38026 Grenoble, France and(2) Department of Physics, University of California, Davis, CA 95616, USA Abstract
We update the constraints on Two-Higgs-Doublet Models of Type I and II discussed inarXiv:1405.3584 using the latest LHC measurements of the ∼ . γγ decaymode moved closer to SM expectations. ∗ [email protected] † [email protected] ‡ [email protected] § [email protected] a r X i v : . [ h e p - ph ] S e p . INTRODUCTION In a recent paper [1], we provided a comprehensive analysis of the status of Two-Higgs-Doublet Models (2HDMs) of Type I and Type II, considering both the cases where theobserved Higgs particle at the LHC is the lighter CP-even state h or the heavier CP-evenstate H . To this end, we performed scans of the 2HDM parameter space taking into accountall relevant constraints from precision electroweak data, from stability, unitarity and pertur-bativity of the potential, as well as from B physics and from the direct searches at LEP. Wealso employed the most recent limits from searches for heavy Higgs-like states at the LHC.The central piece of the analysis however was to check for consistency with the various signalstrength measurements of the observed ∼ . i.e. according tothe status of the Moriond and LHCP 2013 conferences (see [2] for a summary).Since then, a number of new measurements or updates of existing ones were publishedby the experimental collaborations. Most significant, from the point of view of our analysisof the 2HDMs, were the long-awaited final results for the γγ decay mode from CMS [3] inJuly and the update of the γγ results from ATLAS [4] at the end of August 2014. Therewere also several other important new measurements or updates; for example uncertaintieshave been significantly reduced for the fermionic channels, particularly for H → b ¯ b in ttHproduction. All these new results were put together and analyzed in global coupling fitsin [5].In the present note, we now revisit the analysis of [1] and ask what are the implicationsin the 2HDM context of all these new (or updated) results on the signal strengths of the ∼ . χ calculation for the signal strengths at 125 . µ ggF+ttH ( Y ) versus µ VBF+VH ( Y ) planes, Eq. (5) of [1], with the new numbers presented in Ta-ble I of [5]. Points for which χ Y < .
18 for each decay mode Y = γγ, V V (= W W, ZZ ) , b ¯ b, τ τ (that means points that are consistent within 95.4% confidence level (CL) with the observedsignal strengths for each decay mode Y ) and that in addition pass all other relevant con-straints will be called “postLHC8(2014)-FDOK” and compared to the corresponding pointsof [1], called “postLHC8(2013)-FDOK”. In the plots, we will moreover identify the points2hat fit both the 2013 and the 2014 analyses as “postLHC8(2013 & 2014)-FDOK”.We note that in this addendum we focus on pointing out the small modifications that arisefrom the latest ATLAS and CMS results. For a detailed physics discussion and implicationsfor future measurements, e.g. at the next run of the LHC at √ s = 13 −
14 TeV, we refer to themain paper [1], whose results overall remain perfectly valid. For notations and conventions,we also refer to [1].
II. m h ∼ . We begin by showing in Fig. 1 the points surviving all constraints in the cos( β − α )versus tan β plane for the m h ∼ . | cos( β − α ) | get slightlymore constrained, while in Type II there is a narrow strip around cos( β − α ) ≈ − . β (cid:46) β − α ) are allowed from the 2014 measurements. (Such orange points, whichwere not compatible with the 2013 results but are now allowed after the 2014 updates FIG. 1. Constraints in the cos( β − α ) versus tan β plane for m h ∼ . i.e. thestatus considered in [1]), red points are those which remain valid when employing the Summer2014 updates, and orange points are those newly allowed after the Summer 2014 updates. Notethat the latter occur only in Type II but not in Type I models. IG. 2. As in Fig. 1, but for µ hgg ( γγ ) and µ h VBF ( γγ ) signal strengths versus cos( β − α ) (top row), µ h VBF ( γγ ) versus µ hgg ( γγ ) (middle row), and µ hgg ( ZZ ) versus µ hgg ( γγ ) (bottom row). do occur only in Type II but not in Type I.) The banana-shaped branch spanning from(tan β, cos( β − α )) ≈ (3 , .
6) to (40 , .
1) is still present; this corresponds to the solutionwith a flipped sign for C D .The reason for these slight changes lies mostly in the new combined signal strengths for the γγ decay mode: (cid:98) µ ggF+ttH ( γγ ) = 1 . ± .
24 and (cid:98) µ VBF+VH ( γγ ) = 1 . ± .
46 with a correlation4
IG. 3. Reduced couplings C hV vs. C hF in Type I (left plot) and C hU vs. C hD in Type II (right plot)for m h ∼ . of ρ = − .
30 [5], as compared to (cid:98) µ ggF+ttH ( γγ ) = 0 . ± .
28 and (cid:98) µ VBF+VH ( γγ ) = 1 . ± . ρ = − .
38 in 2013 [2]. The result, after combining ATLAS and CMSdata, is that the best-fit signal strength in the ggF mode has increased (although the newcentral value is consistent at the 1 σ level with the 2013 results) while that in VBF+VHproduction has come down by a bit more than 1 σ . The γγ signal strengths as a function ofcos( β − α ) are shown in the top row of Fig. 2. Here, one sees explicitly that the low valuesof µ hgg ( γγ ) ≈ . − . µ hgg ( γγ ) obtainable in Type I, higher values up to about µ hgg ( γγ ) ≈ . ∼ . σ range that should be allowed in principle. The limitationin fact comes from the h → V V (= W W, ZZ ) decay mode in ggF production, for which wehave (cid:98) µ ggF+ttH ( V V ) = 1 . ± .
17 from the 2014 measurements, and hence µ hgg ( ZZ ) < . (cid:98) µ ggF+ttH ( V V ) = 0 . ± .
16 in Spring 2013). The correlationsbetween signal strengths in different channels are illustrated in the middle and bottom rowsof Fig. 2.In Fig. 3, we plot C hV vs. C hF (Type I) and C hU vs. C hD (Type II) for the m h ∼ . h → γγ rate can come from a suppression of C D , which suppresses B ( h → b ¯ b ). In Type I, since C U = C D ≡ C F , this goes hand-in-hand with a reduction of In ttH production, uncertainties are still too large to have any impact. IG. 4. Correlation of BR( h → AA ) and µ hgg ( γγ ) for m h ∼ . m A < m h / the hgg coupling; depending on which effect dominates, µ h VBF ( γγ ) can be enhanced, while µ hgg ( γγ ) is suppressed (cf. the middle-left plot in Fig. 2). This does not occur in Type II,where enhancement/suppression of C U and C D is anti-correlated. In this case C D < C D > µ hgg ( γγ ) and µ h VBF ( γγ ); since C U works in the same direction, the effect can be more pronounced for µ hgg ( γγ ) than for µ h VBF ( γγ ) (cf. the middle-right plot in Fig. 2)Regarding the allowed ranges of m H , m A and m H ± , there is no visible change with respectto [1]. It is however interesting to take a closer look at the region m A < m h /
2, where h → AA decays are possible. In Fig. 4, we show the correlation between B ( h → AA ) and µ hgg ( γγ ). Asone can see, large values of the former imply suppression of the latter. Although the effect issmall, the requirement µ hgg ( γγ ) > .
77 at 2 σ constrains the maximum B ( h → AA ) that canbe obtained in Type I and Type II models. (In Type II, however, a focused scan would beneeded for a quantitative interpretation of the results.) This limit is actually stronger thanthe “direct” constraint on unseen decays, B new < .
22 [5], from the generic C U , C D , C V < III. m H ∼ . Let us now turn to the case that the observed Higgs state at ∼ . H . Analogous to Fig. 1, we show in Fig. 5 the m H ∼ . β − α ) versus tan β plane after all constraints have been applied. As before, weobserve a slight narrowing of the allowed sin( β − α ) range, but no visible change in the6 IG. 5. Constraints in the sin( β − α ) versus tan β plane for m H ∼ . γγ signal strengths versus sin( β − α ). tan β direction. It is interesting to note, however, that in Type I m H ∼ . β (cid:38)
1. It is also remarkable that, while in Type I sin( β − α ) can still vary from about − . . C V (cid:38) . β − α ) being quite rare andassociated with the branch having a negative sign for C HD .The γγ signal strengths as a function of sin( β − α ) are shown in Fig. 6. Correlations ofsignal strengths are illustrated in Fig. 7. Analogous arguments as for the m h ∼ . µ Hgg ( γγ ), µ H VBF ( γγ )and µ Hgg ( ZZ ) in Type II is much stronger than for m h ∼ . µ Hgg ( γγ ) and µ Hgg ( ZZ ) values beyond those found in [1]. Thesewould be removed if future measurements show that µ Hgg ( γγ ) ( µ Hgg ( ZZ )) is within 10% (20%)of unity. As was noted in [1], if ≤ ±
5% deviations from the SM are required for both the ZZ and γγ final states then the upper plots show that only a few points of the Type Imodel having µ Hgg ( γγ ) > ∼ .
95 can survive and that all
Type II points will be removed bythis constraint.As regards the h , A and even the H ± masses associated with a good fit by the H to theLHC data and other limits, there is not much change with respect to [1]. In particular therange of h and A masses discussed in [1] remains valid, the only modification is the slightnarrowing in sin( β − α ) already visible in Fig. 5. IV. CONCLUSIONS
Overall, the new ATLAS and CMS analyses from Summer 2014 lead to relatively minormodifications of the preferred parameter ranges in 2HDM models of Type I and Type II,the most significant changes being slight upward shifts of the central µ gg ( γγ ) and µ gg ( V V )values. In both Type I and Type II this results in the exclusion of points with too low µ gg ( γγ ) and/or µ gg ( V V ). In addition, in Type II points with somewhat higher µ gg ( γγ )and µ gg ( V V ) (beyond those allowed in the 2013 analysis) are now allowed; such new pointshowever do not occur in Type I. Apart from these small shifts the results and in particularthe conclusions of [1] do not change.A possibly important particular point is that the scenarios with low m A <
100 GeVthat escape all LEP and (so far) LHC limits and yet have quite substantial gg → A and bbA production cross sections survive the latest data. It will be interesting to probe thesescenarios, which are possible for both Type I and Type II in the m h ∼ . m H ∼ . IG. 7. Correlations of signal strengths for m H ∼ . future LHC running at higher energy. V. ACKNOWLEDGEMENTS
We thank J´er´emy Bernon for discussions on the new signal strength constraints. Thiswork was supported in part by US DOE grant DE-SC-000999 and by the “Investissementsd’avenir, Labex ENIGMASS”. Y.J. is supported by LHC-TI fellowship US NSF grant PHY-0969510. [1] B. Dumont, J. F. Gunion, Y. Jiang, and S. Kraml, Phys.Rev.
D90 , 035021 (2014),arXiv:1405.3584 [hep-ph].[2] G. Belanger, B. Dumont, U. Ellwanger, J. Gunion, and S. Kraml, Phys.Rev.
D88 , 075008(2013), arXiv:1306.2941 [hep-ph].
3] V. Khachatryan et al. (CMS Collaboration), (2014), arXiv:1407.0558 [hep-ex].[4] G. Aad et al. (ATLAS Collaboration), (2014), arXiv:1408.7084 [hep-ex].[5] J. Bernon, B. Dumont, and S. Kraml, (2014), arXiv:1409.1588 [hep-ph].(ATLAS Collaboration), (2014), arXiv:1408.7084 [hep-ex].[5] J. Bernon, B. Dumont, and S. Kraml, (2014), arXiv:1409.1588 [hep-ph].