Timid composite pseudo-scalars hiding at the LHC?

Thomas Flacke IBS CTPU

KEK-NCTS-KIAS Workshop on PPP and Cosmology,KIAS, Nov. 7th, 2017

based on: A. Belyaev, G. Cacciapaglia, H. Cai, T. Flacke,

H. Serodio, A. Parolini [PRD 94 (2016) no 1, 015004] A. Belyaev, G. Cacciapaglia, H. Cai, G. Ferretti,

T. Flacke, H. Serodio, A. Parolini [JHEP 1701 (2017) 094] G. Cacciapaglia, G. Ferretti,T. Flacke, H. Serodio, [arXiv:1710.11142]

1/13

H• Motivation

• UV embeddings of composite Higgs models and the timid composite pNGB pseudo-scalar (TCP).

• A case study of a search for a light (10-100 GeV) TCP in the di-tau + jet channel.

• Conclusions

??

Outline

An alternative solution to the hierarchy problem: • Generate a scale ΛHC<<Mpl through

a new confining gauge group. • Interpret the Higgs as a pseudo-Nambu-

Goldstone boson (pNGB) of a spontaneously broken global symmetry of the new strong sector.

The price to pay: • From the generic setup, one expects additional

resonances (vectors, vector-like fermions, scalars) around ΛHC (and additional light pNGBs?).

• The non-linear realization of the Higgs yields deviations of the Higgs couplings from their SM values.

• …and many model-building questions …

Running of the new strong coupling

αs

mh

H

ΛHC=g*f~few TV

1019GeV

Mpl

eV

eV

125 GeV“Higgs”

Kaplan, Georgi [1984]

O(few TeV)

f > 800 GeV

f

(𝜓𝜓)

T’

ρ, ρµ

2/13

Motivation for a composite Higgs

Composite Higgs Models: Towards an underlying model

and its low-energy phenomenologyFerretti etal. [JHEP 1403, 077] classified candidate models which: c.f. also Gherghetta etal (2014), Vecchi (2015) for early related works on individual models

• contain no elementary scalars (to not re-introduce a hierarchy problem),

• have a simple hyper-color group, • have a Higgs candidate amongst the pNGBs of the bound states, • have a top-partner amongst its bound states (for top mass via partial

compositeness), • satisfy further “standard” consistency conditions (asymptotic freedom,

no anomalies),

3/13

List of "minimal" CHM UV embeddings

4/13

The resulting models have several common features: • All models require two types of hyper-quarks 𝜓, 𝜒 . The Higgs is realized as

a 𝜓𝜓 bound state. Top partners are realized as 𝜓𝜓𝜒 or 𝜓𝜒𝜒 bound states. • None of the models has the minimal EW coset SO(5)/SO(4). The smallest

EW cosets are instead SU(4)/Sp(4), SU(5)/SO(5), or SU(4)xSU(4)/SU(4). • All models contain two spontaneously broken U(1) symmetries (global

phases of 𝜒, 𝜓), which are singlets under the Standard Model group. One linear combination (𝜂’) is anomalous under the hyper color group (and hence expected to be heavy). The orthogonal combination (a) is an SM singlet which couples to the SM only through the Wess-Zumino-Witten anomaly. Hence, a pNGB with (calculable and fixed) WZW couplings is a genuine prediction of the UV completions under consideration. [PRD 94 1, 015004, JHEP 1701, 094]

5/13

"Minimal" CHM UV embeddings

A timid composite pNGB pseudo-scalar (TCP)

• The mass ma must result from explicit breaking of the U(1) symmetries → treated as free parameter in the effective theory.

• fa results from chiral symmetry breaking. • The WZW coefficients 𝜅i are fully determined by the quantum numbers of 𝜒, 𝜓.

• The coefficient Cf are also fixed in each individual model (up to discrete and few choices).

• Effective couplings of a to the Higgs are induced at loop level :

ALL composite Higgs model embeddings studied contain an SM singlet pseudo scalars a which is described by the effective Lagrangian

CTPU-17-35

Revealing timid pseudo-scalars with taus at the LHC

Giacomo Cacciapaglia,1, 2 Gabriele Ferretti,3 Thomas Flacke,4 and Hugo Serodio5

1Universite de Lyon, France; Universite Lyon 1, Villeurbanne, France

2CNRS/IN2P3, UMR5822, IPNL F-69622 Villeurbanne Cedex, France

3Department of Physics, Chalmers University of Technology, Fysikgarden, 41296 Goteborg, Sweden

4Center for Theoretical Physics of the Universe,

Institute for Basic Science (IBS), Daejeon, 34051, Korea

5Department of Astronomy and Theoretical Physics, Lund University, SE-223 62 Lund, Sweden

(Dated: October 25, 2017)

A light pseudo-scalar that is copiously produced at the LHC may still be allowed by presentsearches. While masses above 65 GeV are e↵ectively covered by di-photon searches, the lower masswindow can be tested by a new search for boosted di-tau resonances. We test this strategy on a setof composite Higgs models with top partial compositeness, where most models can be probed withan integrated luminosity below 300 fb�1.

PACS numbers:

INTRODUCTION

The search for new resonances is one of the mainphysics goals at the LHC, with the discovery of a Higgsboson at an invariant mass of 125 GeV being an illustri-ous example [1, 2]. The e↵orts continue, mainly focusingon high mass objects typically heavier than the Higgs it-self. There are in fact few searches exploring invariantmasses of two Standard Model (SM) particles below, say,100 GeV: one notable case is the search for a di-photonresonance [3, 4], mostly motivated by models that fea-ture an extended Higgs sector, like two Higgs doubletmodels (2HDMs) [5] and the next-to-minimal supersym-metric SM [6].

In this letter, we focus on the LHC phenomenologyof a light new scalar with a mass between 10 and 100GeV which can resonantly decay into a pair of SM par-ticles. Generically, light new scalars are strongly con-strained from electroweak precision measurements (indi-rectly) and from direct searches at LEP and Tevatron. Atthe LHC, besides the above mentioned di-photon chan-nel, light (pseudo)scalars are usually searched for in thedecays of the 125 GeV Higgs boson. This search strat-egy in the 10 to 100 GeV window has been mainly mo-tivated by supersymmetry or 2HDMs. Below roughly10 GeV, strong bounds arise from searches related tomesons, or in experiments looking for light axion-likeparticles (ALPs) [7–10]. Thus, the common lore is thata new scalar, in order to escape detection, needs to beeither very heavy or weakly coupled to the SM.

Note, however, that it is enough to have small cou-plings to electrons and to the electroweak gauge bosonsin order to escape direct LEP searches and electroweakprecision bounds, as well as small couplings to the Higgsto avoid the Higgs portal constraints. Couplings to glu-ons (and heavy quarks) are less constrained, leading to

sizable production rates at the LHC. Candidates of thiskind arise naturally in models of composite Higgs whichenjoy a fermionic UV completion [11–15]. Recent lat-tice results [16] have started to address the mass of suchobject in a specific model [17].In this letter, we will consider this class of models to

explore the 10 to 100 GeV mass window and show that itis, in fact, very poorly tested. A timid composite pseudo-scalar (TCP) arises as the pseudo-Nambu-Goldstone bo-son associated with an anomaly-free U(1) global symme-try in all models of partial compositeness that enjoy a UVcompletion, as defined in Ref. [12]. All the possible mod-els can be classified, and give precise predictions for theproperties of the TCP candidate [15], thus mapping out acomplete landscape of possibilities. We show that, whilesome models are already partly tested by the low massdi-photon searches, others are completely unconstrained.We point out that searches for di-tau resonances (whichnow start at 90 GeV invariant mass at the LHC [18, 19])give very promising signals and could be a powerful com-plementary probe to the di-photon channel, or even bethe only way to access this class of TCPs.

DESCRIPTION OF THE MODELS

The e↵ective Lagrangian we consider is the SM La-grangian augmented by the following terms, up to di-mension five operators:

L =1

2(@µa)(@

µa)� 1

2m2

aa2 �

X

f

iCfmf

faa f�

5 f (1)

+g2sKga

16⇡2faGa

µ⌫Gaµ⌫+

g2KWa

16⇡2faW i

µ⌫Wiµ⌫+

g02KBa

16⇡2faBµ⌫B

µ⌫ .

A pseudo-scalar a described by this general Lagrangianarises, for example, in UV completions of compos-

2

ite Higgs models which were classified and studied inRefs [12, 15]. Within this class of models, the cou-pling to the SM fermions given in Eq. (1) is only thefirst term of the expansion of the spurion coupling�mf (h) eiCfa/fa L R + h.c. (generating the fermionsmasses), which breaks explicitly the U(1) shift symme-try. A derivative coupling of the TCP to fermions ofthe form (@µa/fa) f�5�µ f is absent in these modelssince the SM fermions are neutral under the TCP U(1)charge. Although such a coupling can be obtained by us-ing the fermion equations of motion and integrating bypart the leading term given in Eq. (1), the two couplingsare of genuinely di↵erent origin [20], as manifested in thehigher-order expansion of the spurion coupling. Startingfrom the complete spurion coupling, a coupling of theHiggs to two TCPs, as well as to one TCP and Z boson,arise at loop level and are given by

Lhaa =3C2

t m2

tt8⇡2f2

avlog

⇤2

m2

t

h(@µa)(@µa), (2)

LhZa =3Ctm2

t gA2⇡2fav

(t � V ) log⇤2

m2

t

h(@µa)Zµ, (3)

where we list only the e↵ect of the log-divergence (⇤ ⇠4⇡fa), gA = �g/(4 cos ✓W ) is the axial coupling of the Zto tops, and V,t are the corrections from compositenessto the coupling of the Higgs to vectors and tops, respec-tively. As X = 1 + O(v2/f2

a ), our result agrees withthe fact that the only non-zero contribution to the hZacoupling arises from a dimension 7 operator [21].

The couplings to gauge bosons in Eq. (1) arise asanomalous couplings if the TCP is a (SM singlet) boundstate of underlying SM charged hyperfermions. In thiscase, the anomaly coe�cients Kg,W,B are fully deter-mined by the charges of the hyperfermions. We refer to[15] for an extensive description of a classification of UVcompletions giving rise to this TCP, which yields twelvemodels. For the purpose of this letter, the TCP dynamicsin the twelve models is fully specified by the numericalcouplings in Table I 1. Note that, due to the small TCPmass, top loops also give additional sizable contributionsto the couplings to gauge bosons (not included in thetable, but included in our analysis). Our goal is to con-front the TCP with the existing searches and to proposenew, more sensitive searches for such object. We treatthe mass ma and the decay constant fa of the TCP asfree parameters. In composite Higgs UV completions,fa is related to the composite Higgs decay constant f ,entering in the usual alignment parameter ⇠ = v2/f2

,by a relative coe�cient that was estimated in [15] andis summarized in Table I. Since bounds on composite

1

The model in [17] is denoted by M6 in this work and in [15],

while the model [11] is denoted by M8.

Kg KW KB Cf fa/f M1 -7.2 7.6 2.8 2.2 2.1M2 -8.7 12. 5.9 2.6 2.4M3 -6.3 8.7 -8.2 2.2 2.8M4 -11. 12. -17. 1.5 2.0M5 -4.9 3.6 0.40 1.5 1.4M6 -4.9 4.4 1.1 1.5 1.4M7 -8.7 13. 7.3 2.6 2.4M8 -1.6 1.9 -2.3 1.9 2.8M9 -10. 5.6 -22. 0.70 1.2M10 -9.4 5.6 -19. 0.70 1.5M11 -3.3 3.3 -5.5 1.7 3.1M12 -4.1 4.6 -6.3 1.8 2.6

TABLE I: Couplings in the twelve minimal composite HiggsUV embedding models [15] which are used as benchmark mod-els. For the top, several possibilities arise depending on thechoice of top partner representation: here, as an illustration,we take the same coupling as for lighter fermions, whose massarise from bilinear four-fermion interactions. The ratio fa/f indicates the expected scale-ratio of the TCP decay constantfa and the composite Higgs decay constant f .

Higgs models require f & 800 GeV, fa is expected tobe naturally of the order of 1÷ 2 TeV.

BOUNDS FROM EXISTING SEARCHES

M1M2M3M4M5

M6M7M8M9M10

M11M12

10 20 30 40 50 60 70 80 90 100500

1000

20003000

5000

10000

Ma @GeVD

f a@Ge

VD

FIG. 1: Constraints on fa as a function of ma for the bench-mark models M1 - M12, defined in Table I. The bounds arisefrom di-muon searches [22, 23] in the low mass range, di-photon searches [3, 4] in the higher one and from the BSMdecay width of the Higgs [24] below 65 GeV.

Since the TCP is a gauge singlet, its couplings to Zand W are induced by the anomaly and top loops, thusthey are always much smaller than those of a SM Higgsboson. Hence, bounds from all LEP searches for a lightHiggs, which are based on Z associated production, areevaded. At hadron colliders the TCP can be is copi-ously produced via gluon fusion. However, only veryfew Tevatron or LHC two-body resonant searches reach

6/13

• a is produced in gluon fusion (controlled by Kg/fa). • Assoc. production with a Z is tiny ➝ No bounds from LEP Higgs searches. • a decays to gg, WW, ZZ, Z𝛾, 𝛾𝛾, ff with fully determined branching ratios. • For heavier a, LHC di-boson searches apply [JHEP 1701, 094]. • For light a (translating existing bounds and searches):

TCP Phenomenology

𝜇𝜇 [PRL109, 121801]

(CMS) [ATLAS-CONF-2011-020]

𝛾𝛾 [PRL113, 17801]

(ATLAS) [CMS-PAS-HIG-17-013]

BR(h➝BSM)<.34 [JHEP1608, 045] (ATLAS+CMS)

7/13

How can we search the gap at low mass?

The models are poorly constrained in the mass range 15 - 65 GeV. • Weak indirect bounds from h➝aa (BSM) which will not

dramatically increase as the bound on fa scales with Br(h➝aa)1/4 . • h➝aa➝ 4𝛾 , bb𝜇𝜇, bb𝜏𝜏, etc. have very low signal rate due to small

haa coupling and small a➝𝛾𝛾,ff branching ratios. • same applies to h➝Za. • b-associated production is small. • t-associated production could yield bounds in future searches.

[EPJC 75, 498]

• Extending high resolution 𝜇𝜇 resonance searches to higher mass? • Extending 𝛾𝛾 resonance searches to even lower mass? • …or looking for other decay channels: 𝜏𝜏!

8/13

How can we search the gap at low mass? 𝜏𝜏!

The gluon-fusion production cross section for light a is large…

… and the 𝜏𝜏 branching ratio is (for most models) not small.

9/13

How can we search the gap at low mass? 𝜏𝜏!Soft 𝜏lep or 𝜏had cannot be used to trigger on, but initial state radiation can boost the gg ➝ a ➝ 𝜏𝜏 system (at the cost of production cross section, but we have enough). As a very naive proof of principle analysis we look for a j 𝜏𝜇 𝜏e final state (jet + opposite sign, opposite flavor leptons) with cuts: • pT𝜇 > 42 GeV (for triggering) • pTe > 10 GeV • ΔR𝜇j > 0.5, ΔRej > 0.5, • ΔR𝜇e < 1.0 • no lower cut on ΔR𝜇e ! • m𝜇e > 100 GeV

Main background: Z/𝛾*+jets: 35 fb, tt+jets: 70 fb, Wt+jets: 7.4 fb, VV+jets: 13 fb.

10/13

How can we search the gap at low mass? 𝜏𝜏!

11/13

Resulting projected reach for 300 fb-1

How can we search the gap at low mass? 𝜏𝜏!

Note: This first proof of principle study is highly non-optimized. • Cutting harder on ΔR𝜇e can

substantially increase background suppression for the lighter mass range.

• We did not use any 𝜏 ID or triggers.

• We only used the OSOF lepton channel. 𝜏𝜇𝜏𝜇, 𝜏𝜇𝜏had, 𝜏had𝜏had have larger branching ratios but require a more careful background analysis. [And needs tagging efficiencies for boosted 𝜏𝜇𝜏had, 𝜏had𝜏had systems which are beyond our capabilities, but possible for experimentalists.]

12/13

• A light pseudo-scalar a is predicted in a large class of Composite Higgs UV embeddings. The models are highly predictive because the branching ratios of different di-boson and di-lepton channels are fully determined by the quantum numbers of the underlying fermion field content.

• In a mass range of 15 - 65 GeV, to our knowledge, none of the existing LEP, Tevatron, and LHC searches are sensitive to this pseudo-scalar. The timid composite pseudo-scalar (TCP) hides well.

• We propose to search for the TCP in the di-tau channel with ISR against which the di-tau system recoils.

• Our (very naive and conservative) initial study in the jet + OSOF lepton final state shows very good sensitivity in the 15 - 65 GeV mass range.

13/13

Conclusions

Backup Slides