Exotic decays of the 125 GeV Higgs boson

We perform an extensive survey of nonstandard Higgs decays that are consistent with the 125 GeV Higgs-like resonance. Our aim is to motivate a large set of new experimental analyses on the existing and forthcoming data from the Large Hadron Collider (LHC). The explicit search for exotic Higgs decays presents a largely untapped discovery opportunity for the LHC collaborations, as such decays may be easily missed by other searches. We emphasize that the Higgs is uniquely sensitive to the potential existence of new weakly coupled particles and provide a unified discussion of a large class of both simplified and complete models that give rise to characteristic patterns of exotic Higgs decays. We assess the status of exotic Higgs decays after LHC run I. In many cases we are able to set new nontrivial constraints by reinterpreting existing experimental analyses. We point out that improvements are possible with dedicated analyses and perform some preliminary collider studies. We prioritize the analyses according to their theoretical motivation and their experimental feasibility. This document is accompanied by a Web site that will be continuously updated with further information [http://exotichiggs.physics.sunysb.edu].

Figure 1

Sensitivity of a 125 GeV Higgs to light weakly coupled particles. Left: Exotic Higgs branching fraction to a singlet scalar $s$ versus the singlet’s mass $m_s$, assuming the interaction Eq. (1) is solely responsible for the $h\to ss$ decay. If the interaction in Eq. (1) generates the $s$ mass, the result is the orange curve; the other curves are for fixed and independent values of $\zeta$ and $m_s$. Right: Exotic Higgs branching fraction to a new fermion $\psi$ interacting with the Higgs as in Eq. (2) to illustrate the sensitivity of exotic Higgs decay searches to high scales, here $\mathrm{\Lambda }$. We take here $\mu =m_{\psi }$.

Figure 2

The exotic Higgs decay topologies we consider in this document, along with the labels we use to refer to them. Every intermediate line in these diagrams represents an on-shell, neutral particle, which is either a $Z$ boson or a BSM particle.

Figure 3

Size of the cubic coupling ${\mu }_v$ in units of Higgs expectation value $v$ to yield the indicated $h\to ss$ branching fraction as a function of singlet mass, as given by Eq. (8).

Figure 4

Left: Branching ratios of a $CP$-even scalar singlet to SM particles, as a function of $m_s$. Right: Branching ratios of exotic decays of the 125 GeV Higgs boson as a function of $m_s$, in the $\mathrm{SM}+\text{scalar}$ model described in the text, scaled to $\mathrm{Br}(h\to ss)=1$. Hadronization effects likely invalidate our simple calculation in the shaded regions.

Figure 5

Required mixing angle between the doublet and singlet-sector pseudoscalar for $\mathrm{Br}(h\to aZ)=10\%$, assuming no other exotic Higgs decays and $\alpha =\pi /2-\beta$ (decoupling limit).

Figure 6

Branching ratios of a singletlike pseudoscalar in the $2\mathrm{HDM}+\mathrm{S}$ for type-I Yukawa couplings. Decays to quarkonia likely invalidate our simple calculations in the shaded regions.

Figure 7

Branching ratios of a singletlike pseudoscalar in the $2\mathrm{HDM}+\mathrm{S}$ for type-II Yukawa couplings. Decays to quarkonia likely invalidate our simple calculations in the shaded regions.

Figure 8

Branching ratios of a singletlike pseudoscalar in the $2\mathrm{HDM}+\mathrm{S}$ for type-III Yukawa couplings. Decays to quarkonia likely invalidate our simple calculations in the shaded regions.

Figure 9

Branching ratios of a singletlike pseudoscalar in the $2\mathrm{HDM}+\mathrm{S}$ for type-IV Yukawa couplings. Decays to quarkonia likely invalidate our simple calculations in the shaded regions.

Figure 10

Singlet scalar branching ratios in the $2\mathrm{HDM}+\mathrm{S}$ for different $\tan \beta ,{\alpha }^{\prime }$ and Yukawa coupling type. These examples illustrate the possible qualitative differences to the pseudoscalar case, such as dominance of $s\to c\overline{c}$ decay above $b\overline{b}$ threshold; democratic decay to $b\overline{b}$ and ${\tau }^{+}{\tau }^{-}$; and democratic decay to $c\overline{c}$ and ${\tau }^{+}{\tau }^{-}$. Hadronization effects likely invalidate our simple calculations in the shaded regions.

Figure 11

Higgs branching fraction into Majorana fermions $\chi$ resulting from the partial width of Eq. (30), as a function of the Higgs portal scale $M$ and the mass of the fermion $m_{\chi }$. We fix the coupling $\kappa$ to be equal to 1.

Figure 12

Constraints on $\epsilon ,m_{Z_D}$ for pure kinetic mixing (no additional source of $Z\text{−}{Z}_D$ mass mixing) for $m_{Z_D}\sim \mathrm{MeV}–10\mathrm{GeV}$. The black dashed line separates prompt ($c\tau <1\mu \mathrm{m}$) from nonprompt $Z_D$ decays. The three blue lines are contours of $\mathrm{Br}(h\to Z{Z}_D)$ of ${10}^{-4}$, ${10}^{-5}$, and ${10}^{-6}$, respectively. Shaded regions are existing experimental constraints [140, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 167, 168, 169, 170, 171, 172, 173, 174]; see, e.g., [134] for a recent review. The red shaded region “CMS” is a new limit we derived by recasting the CMS $20+5{\mathrm{fb}}^{-1}$ $h\to Z{Z}^{*}$ analysis [175], as described in Sec. 10. (We obtain a similar bound from the corresponding ATLAS analysis [176].) This new bound can be optimized with a dedicated LHC measurement, likely improving upon the electroweak precision measurement bounds (green region labeled “EWPM” [167]) for some masses.

Figure 13

(a) Branching ratios for $Z_D$ decay, to lowest order and without QCD corrections, assuming decays to the dark sector are kinematically forbidden. Hadronization effects likely invalidate our simple calculation in the shaded region. (b) Branching ratios for $Z_D$ decay for $m_{Z_D}\lesssim 3\mathrm{GeV}$, including nonperturbative QCD effects.

Figure 14

The dominant exotic Higgs decays in the $\mathrm{SM}+\mathrm{V}$ model. The $h\to Z{Z}_D$ matrix element is proportional to the gauge kinetic mixing $\epsilon$, while $h\to Z_D{Z}_D$ and $h\to ss$ are controlled by the Higgs mixing parameter $\zeta$. The vertex $hs{Z}_D$ is present but suppressed by both mixings.

Figure 15

Left: Mass plane in the $\mathrm{SM}+\mathrm{V}$ model with different exotic Higgs decays for $\zeta \gg \epsilon$ (i.e., when the mixing between the Higgs and dark Higgs dominates over the kinetic mixing). The black contours are the values of $\zeta \times {10}^3$ required for $\mathrm{Br}(h\to Z_D{Z}_D,ss)=10\%$. Region $A$ is the case examined by [177] (the dotted red line indicates $m_h=m_s$). Region $C$ has no exotic Higgs decays. Region $D$ reproduces the $\mathrm{SM}+\mathrm{S}$ model of Sec. 1c1. Region $B$ has both $h\to ss$ and $h\to Z_D{Z}_D$ decays, with the $h\to ss$ fraction of exotic decays shown on the right. In the upper left shaded region, $s\to Z_D{Z}_D$ is the dominant decay mode of the dark scalar. This allows the Higgs to decay to up to eight SM fermions.

Figure 16

Branching ratios of the Higgs into neutralinos: $\mathrm{Br}(h\to {\chi }_1{\chi }_1)$ and $\mathrm{Br}(h\to {\chi }_1{\chi }_2)$ are shown in blue and red, respectively. The yellow region is the region excluded by the LEP bound on the $Z$ invisible width. The region below the dashed green line is the region with a lightest chargino below the LEP bound of $\sim 100\mathrm{GeV}$. The input parameters are $\tan \beta =10$ and $M_2=300\mathrm{GeV}$ (a), $M_2=150\mathrm{GeV}$ (b).

Figure 17

Two significant fermionic decay topologies of the SM-like Higgs boson in the $PQ$-symmetry limit. (a) Depending on whether $\min \{m_s,m_a\}$ exceeds $m_{{\chi }_2}-m_{{\chi }_1}$, $a(s)$ may or may not be on shell. (b) To be non-negligible, the radiative ${\chi }_2$ decay requires $\min \{m_s,m_a\}>m_{{\chi }_2}-m_{{\chi }_1}$.

Figure 18

Expected signal strength observed in a $h\to 2b$ search, assuming SM-like Higgs production and couplings with the exception of a new $h\to 2a\to 4b$ decay mode with selection efficiency $r_{4b}$ relative to the efficiency of SM-like $h\to 2b$ events.

Figure 19

Dimuon invariant mass spectrum, $m_{\mu \mu }$, for signal ($m_a=30\mathrm{GeV}$) and backgrounds for $25{\mathrm{fb}}^{-1}$ at 8 TeV LHC after all kinematic cuts (except for $m_{\mu \mu }$ cuts) with (left) no $b$ tag, (middle) at least one $b$ tag, and (right) two $b$ tags. For the signal normalization, we assume $\mathrm{Br}(h\to aa)=10\%$ and $2\times \mathrm{Br}(a\to b\overline{b})\mathrm{Br}(a\to {\mu }^{+}{\mu }^{-})=1.7\times {10}^{-3}$ as in Table 3.

Figure 20

Expected 95% C.L. sensitivity to $\mathrm{Br}(h\to aa\to b\overline{b}{\mu }^{+}{\mu }^{-})$ for $25{\mathrm{fb}}^{-1}$ data at 8 TeV LHC. The solid, dashed, and dotted lines show the limits for at least one $b$ tag, two $b$ tags, and no $b$ tag, respectively.

Figure 21

Expected 95% C.L. sensitivity to $\mathrm{Br}(h\to aa)$ from a $h\to b\overline{b}{\mu }^{+}{\mu }^{-}$ search as a function of $\mathrm{Br}(a\to b\overline{b})$ and $\mathrm{Br}(a\to {\tau }^{+}{\tau }^{-})$, assuming that the pseudoscalar coupling to leptons is proportional to the lepton masses. We show $m_a=30\mathrm{GeV}$ (left) and $m_a=60\mathrm{GeV}$ (right) with $25{\mathrm{fb}}^{-1}$ of data at the 8 TeV LHC (see text for further details). The red solid lines and blue dashed lines present the limits for at least one $b$ tag and two $b$ tags, respectively. The corresponding sensitivities to $\mathrm{Br}(h\to aa\to b\overline{b}{\mu }^{+}{\mu }^{-})$ are given in Fig. 20.

Figure 22

Expected 95% C.L. sensitivity to $\mathrm{Br}(h\to aa\to b\overline{b}{\mu }^{+}{\mu }^{-})$ for $100{\mathrm{fb}}^{-1}$ of data at 14 TeV LHC. The solid, dashed, and dotted lines show the limits for at least one $b$ tag, two $b$ tags, and no $b$ tag, respectively.

Figure 23

Expected 95% C.L. sensitivity to $\mathrm{Br}(h\to aa)$ from a $h\to b\overline{b}{\mu }^{+}{\mu }^{-}$ search as a function of $\mathrm{Br}(a\to b\overline{b})$ and $\mathrm{Br}(a\to {\tau }^{+}{\tau }^{-})$, assuming that the pseudoscalar coupling to leptons is proportional to the lepton masses. We show $m_a=30\mathrm{GeV}$ (left) and $m_a=60\mathrm{GeV}$ (right) with $100{\mathrm{fb}}^{-1}$ of data at the 14 TeV LHC (see text for further details). The red solid lines and blue dashed lines present the limits for at least one $b$ tag and two $b$ tags, respectively. The corresponding expected sensitivities to $\mathrm{Br}(h\to aa\to b\overline{b}{\mu }^{+}{\mu }^{-})$ are given in Fig. 20.

Figure 24

Estimated exclusion of $\mathrm{Br}(h\to 2a\to 4\tau )$ from LHC multilepton and same-sign dilepton searches: (1) CMS three-lepton from [298] in red, (2) CMS four-lepton from [300] in blue, (3) ATLAS same-sign dilepton from [302] in green. The black line shows a combination of the multilepton searches (1) and (2). [The combination of all channels, including (3), is less constraining by several percent.]

Figure 25

Median estimates of expected indirect exclusions on $\mathrm{Br}(h\to 2a\to 4\tau )$ using the subdominant $(a\to 2\tau )(a\to 2\mu )$ channel and exploiting that leptonic branching fractions of $a$ are mass ordered. The results are based on a simulated ${\mu }^{+}{\mu }^{-}$ resonance search in $\ge 3\ell$ events, assuming the 2012 data set. Since we cannot reliably predict the background under the resonance peak, we show expected exclusions for $B=0$, 5, 10 and 20 events, respectively. We neglect systematic uncertainties. (The lowest displayed mass is 4.0 GeV.)

Figure 26

Projected 95% C.L. limits on the branching fraction for $h\to 2\gamma 2g$ in associated production ($Wh$), as a function of $m_s$. The blue curves refer to $300{\mathrm{fb}}^{-1}$ (solid) and $100{\mathrm{fb}}^{-1}$ (dashed), both at the 14 TeV LHC. The dashed-dotted green curve shows a conservative estimate of the sensitivity for $20{\mathrm{fb}}^{-1}$ at 8 TeV. All three limits build on the proposed search in [312] ($300{\mathrm{fb}}^{-1}$ at 14 TeV LHC), by scaling background with luminosity but not changing its cross section, while signal is rescaled according to both luminosity and cross section. This underestimates the achievable 8 TeV limit. See text for more details.

Figure 27

Left: 95% C.L. exclusion limit on $\mathrm{Br}(h\to ZX)\times \mathrm{Br}(X\to \ell \ell )$ for $X=Z_D,a$, extracted from the SM $h\to 4\ell$ searches ($\ell =e,\mu$) assuming SM Higgs production rate and ${\mathrm{\Gamma }}_X\ll 1\mathrm{GeV}$. (The lighter dashed lines indicate the expected limit. The large fluctuations in the observed limit are a consequence of low statistics in each bin.) Right: The CMS distribution of $m_{Z2}$ from [175], overlaid with a 23 GeV $h\to ZX\to 4\ell$ signal.

Figure 28

Top left and right: Distribution of lepton pair invariant masses in $4e$ and $4\mu$ events according to the event selection and reconstruction criteria of [175]. The maximum cross section [taking $\mathrm{Br}(h\to Z_D{Z}_D)=1$] in any $2.5\times 2.5\mathrm{GeV}$ square is indicated in each plot to establish a scale. Left: With $m_{Z_D}=20\mathrm{GeV}$, only mispaired $4e$ and $4\mu$ events pass the event selection criteria. Right: With $m_{Z_D}=40\mathrm{GeV}$, both mispaired and correctly paired events are evident, with accumulation of events at the mass of the vector boson visible on the far left edge of the plot. (In this case, $2e2\mu$ events, not shown, also pass the selection criteria and accumulate at the mass of the vector boson.) Bottom: Expected distribution of lepton pair invariant masses for $h\to Z{Z}^{*}\to 4\ell$ with $m_{4\ell }\in (121.5,130.5)$, overlaid with observed 7 and 8 TeV events, from [175].

Figure 29

Left: Distribution of lepton pair invariant masses for signal with $m_{Z_D}=25\mathrm{GeV}$ for all flavor combinations, according to the event selection and reconstruction criteria of [341]. Correctly paired events are shown in blue and make up 55% of the accepted events, while mispaired events, in purple, make up the remaining 45%. Right: Distribution of selected lepton pair invariant masses, from [341]. Note that the scales of the axes differ in the two plots.

Figure 30

Estimated 95% C.L. limits on the branching fraction $\mathrm{Br}(h\to Z_D{Z}_D)$ coming from CMS $h\to Z{Z}^{*}$ [175] (red, dotted line) and ATLAS $ZZ$ cross section [341] (blue, dashed line) measurements. Note that, as seen in Fig. 13 in Sec. 1c5, for this range of $m_{Z_D}$, $\mathrm{Br}(Z_D\to {\ell }^{+}{\ell }^{-})\simeq 0.3$ which implies that $\mathrm{Br}(h\to Z_D{Z}_D\to 4\ell )\simeq 0.09\times \mathrm{Br}(h\to Z_D{Z}_D)$.

Figure 31

Approximate 95% C.L. upper limit on $(\sigma /{\sigma }_{\mathrm{SM}})\times \mathrm{Br}(h\to {\chi }_1{\chi }_2\to \gamma +{\cancel{E}}_T)$ from the results of Ref. [359], for $m_{{\chi }_1}=(0,20,40\mathrm{GeV})<m_{{\chi }_2}$. Solid lines correspond to 100% photon efficiency, and dashed lines to a (flat) 80% photon efficiency.

Figure 32

Approximate 95% C.L. upper limit on $(\sigma /{\sigma }_{\mathrm{SM}})\times \mathrm{Br}(h\to {\chi }_2{\chi }_2\to 2\gamma +{\cancel{E}}_T)$ from the results of Ref. [359], for multiple values of $m_{{\chi }_2}$ as indicated by the text labeling the different curves. Solid lines correspond to 100% photon efficiency, and dashed lines to a (flat) 80% photon efficiency.

Figure 33

Approximate 95% C.L. upper limit on $(\sigma /{\sigma }_{\mathrm{SM}})\times \text{Br}(h\to 2\gamma +{\cancel{E}}_T)$ from the $2\gamma +{\cancel{E}}_T$ search in [359]. The solid lines correspond to 100% photon efficiency, and the dashed lines to a (flat) 80% photon efficiency. Left: Resonant case, where $h\to aa$, one $a$ decays to $\gamma \gamma$ and the other decays invisibly. Right: Cascade case, where $h\to {\chi }_1{\chi }_2$, ${\chi }_2\to s{\chi }_1$, $s\to \gamma \gamma$. Here $m_{{\chi }_1}=0$ and $m_{{\chi }_2}=60\mathrm{GeV}$ (although the limit is insensitive to the particular value of $m_{{\chi }_2}$ as long as it is kinematically allowed).

Figure 34

Unit-normalized distributions of $m_T(2\ell ,{\cancel{E}}_T)$. The blue dashed line shows the ATLAS prediction for SM $h\to W{W}^{*}$ events passing all selection criteria in both 7 and 8 TeV data sets [376]. The purple dotted line shows the distribution for the BSM $h\to 2\ell +{\cancel{E}}_T$ events arising from $h\to {\chi }_2{\chi }_1$ at the 8 TeV LHC in the benchmark model described in the text.

Figure 35

Approximate bounds on the branching fraction for $h\to {\chi }_2{\chi }_2$, assuming (left) $\mathrm{Br}({\chi }_2\to a{\chi }_1)=1$, and (right) $\mathrm{Br}({\chi }_2\to Z_D{\chi }_1)=1$, as a function of $m_{{\chi }_1}$, from [339]. Here solid lines indicate $m_{{\chi }_2}=50\mathrm{GeV}$ and dotted lines $m_{{\chi }_2}=60\mathrm{GeV}$, while red, green, and blue correspond to $m_{a,Z_D}=3$, 1, and 0.4 GeV, respectively. We use tree-level results for $\mathrm{Br}(Z_D\to \mu \mu )$ (see Fig. 13) and a reference $\mathrm{Br}(a\to \mu \mu )=0.1$ (which can occur in type-IV $2\mathrm{HDM}+\mathrm{S}$ models; see Fig. 9).

Figure 36

Higgs phenomenology in the $PQ$-symmetry limit of the NMSSM, as discussed in Sec. 1c8 [54]. Top row: Masses of $s$, $a$, and ${\chi }_1$, respectively. Second and third rows: Branching ratios of the SM-like Higgs $h$ (denoted here as $h_2$) to $b\overline{b}$, $s\overline{s}$, $aa$, ${\chi }_1{\chi }_1$, ${\chi }_1{\chi }_2$, and ${\chi }_2{\chi }_2$, respectively. Bottom row: Branching ratios of the next-to-lightest neutralino ${\chi }_2$ to on-shell ${\chi }_{1}s+{\chi }_{1}a$ and ${\chi }_{1}Z$, respectively. All points are required to have a mass 124–126 GeV for their SM-like Higgs boson. Green (light gray) points are sampled in the ranges $3\le \tan \beta \le 30$, $0.015\le \lambda \le 0.5$, $0.0005\le \kappa \le 0.05$, $-0.8\le {ϵ}^{\prime }\le 0.8$, $-50\mathrm{GeV}\le A_{\kappa }\le 0$, and $0.1\mathrm{TeV}\le \lambda v_s\le 1\mathrm{TeV}$, green (light gray) points cover the whole scan range, red (medium gray) points correspond to the subset satisfying $\lambda <0.30$, $\kappa /\lambda <0.05$ and $\lambda v_s<350\mathrm{GeV}$, while blue (dark gray) points satisfy $\lambda <0.15$, $\kappa /\lambda <0.03$ and $\lambda v_s<250\mathrm{GeV}$.