LHC Potential for ZH ll + invisible Search€¦ · • The ﬁrst direct search for the invisible Higgs at the LHC. BR(H→inv)

LHC Potential for ZH→ll + invisible Search

Snowmass: Seattle Energy Frontier WorkshopJune 30 - July 3, 2013

Hideki Okawa1, Josh Kunkle2, Elliot Lipeles21Brookhaven National Laboratory, 2University of Pennsylvania

ZH(→ll+inv.) Search @ LHC

Hideki Okawa 2Snowmass: Seattle Energy Frontier Workshop, July 2, 2013

• DIRECT search for the invisible decays of Higgs; BSM process

• Cut-based analysis using Z+ETmiss final state.

• Has one of the highest sensitivities among direct H(→inv) search channels. (cf. other channels; VBF, monojet, W+ETmiss)

• LEP apparently has no sensitivity for mH > ~120 GeV

• Complementary approach to the Higgs coupling studies

LEP Higgs WG, LHWG Note 2001-06

ZH→ll invisible

ATLAS-CONF-2013-034

3

ATLAS Moriond Results

Hideki Okawa

• First results at Moriond EW and onwards using the full 2011 & 2012 HCP dataset (4.7 fb-1 + 13.0 fb-1)

• The first direct search for the invisible Higgs at the LHC. BR(H→inv)<0.65@95% CL obs.

• Also showed interpretations for “another” Higgs scenario

[GeV]Hm120 140 160 180 200 220 240 260 280 300

ll in

v) [f

b]A

BR(Z

H×

ZHm

95%

CL

limit:

10

20

30

40

50

60

70 ll inv)ABR(ZH×ZH,SMm

Observed

Expected

m1±

m2±

PreliminaryATLASll(inv)AZH

-1 Ldt=4.7fb0=7TeV, s-1 Ldt=13.0fb0=8TeV, s

Snowmass: Seattle Energy Frontier Workshop, July 2, 2013

Samples for Snowmass Studies

Hideki Okawa 4

• Used Delphes samples produced by the Snowmass production team (Thank you so much)

• Background:

• ZZ/WZ/WW(&Zγ,Wγ): BB samples with 5 HT slices

• Top: ttbar (tt samples) with 5 HT slices, single top (tj, tB samples) to be added for the next round (though almost negligible for this channel)

• Z/W+jets: B samples with inclusive HT production

• Signals:

• ZH→ll+inv. signals: produced with MadGraph5 & ran Delphes v.3.0.9 with different pileup conditions (0, 50, 140)


ATLAS Event Selection

Hideki Okawa 5

• 2 opposite-sign lepton w/ 76 < Mll < 106 GeV; 3rd lepton veto (pT>7 GeV)

• ETmiss > 90 GeV

• Fractional pT difference ( |ETmiss - pTll| / pTll ) < 0.2

• dϕ(ETmiss,ETmiss,trk) < 0.2

• dϕ(l,l) < 1.7

• dϕ(Z, ETmiss) > 2.6

• Jet veto (pT > 20 GeV, |η| < 2.5)

Even

ts /

5 G

eV

110

210

310

410

510

610

710-1 L dt = 13.0 fb0Data

ZMultijetWTopWWWZZZ

(*) WWA, H (*) ZZASM H = 125 GeV)

HSignal (SM ZH, m

ATLAS = 8 TeVsPreliminary ll+invisibleAZH

[GeV]missTE

0 50 100 150 200 250 300Dat

a / M

C

0.60.811.21.4

ATLAS event selection used for Moriond

/32

rad

/Ev

ents

/

10

210

310

410 ZMultijetWTopWWWZZZ

(*) WWA, H (*) ZZASM H = 125 GeV)

HSignal (SM ZH, m

-1 L dt = 13.0 fb0Data

ATLAS = 8 TeVsPreliminary ll+invisibleAZH

) [rad] missT

p,missT

(Eq6

0 0.5 1 1.5 2 2.5 3Dat

a / M

C

00.5

11.5

2

ATLAS-CONF-2013-011

Snowmass: Seattle Energy Frontier Workshop, July 2, 2013ETmiss-related variables

Hideki Okawa

Snowmass Scenarios

6

• 14 TeV scenarios: 300 fb-1 (pileup μ=50) & 3000 fb-1 (pileup μ=140)

• Made minor modifications to the ATLAS ZH(→ll+invisible) event selection for the following reasons

• Missing ET: degradation of resolution due to more pileup

• Removed dΦ(ETmiss, track pTmiss) cut for now. Detailed investigations are needed for the tracks in Delphes samples.

• Jet veto threshold: pileup subtraction is not applied in the Delphes samples (which is different from the ATLAS conditions). So, we simply raised the pT threshold for now.


Hideki Okawa

Snowmass Event Selection

7

[GeV]missTE

0 50 100 150 200 250 300 350 400 450 500

Even

ts

1

210

410

610

810

1010

1210=14 TeVsDelphes

invisible-µ+µ/-e+ eAZH =50µ, -1 L = 300 fb0

V+jetsTopVV

= 125 GeV,H

Signal (m inv) = 1)ABR(H

[GeV]missTE

0 50 100 150 200 250 300 350 400 450 500

Even

ts

1

210

410

610

810

1010

1210=14 TeVsDelphes

invisible-µ+µ/-e+ eAZH =140µ, -1 L = 3000 fb0

V+jetsTopVV

= 125 GeV,H

Signal (m inv) = 1)ABR(H

• ETmiss > 90 → 100 GeV

• |ETmiss - pTll| / pTll < 0.2 → 0.4

• Jet veto pT threshold : 20→45 GeV

14 TeV 300 fb-1 (μ=50) 14 TeV 3000 fb-1 (μ=140)

Changes to the cut thresholds


• ETmiss > 90 → 170 GeV

• |ETmiss - pTll| / pTll < 0.2 → 0.6

• dϕ(l,l) < 1.7 → 0.8

• Jet veto pT threshold : 20→60 GeV

Changes to the cut thresholds

Preliminary

Signal significance without BG uncertainty ~ 1.6 (3.1) for signals w/ BR(H→inv)=10% (20%) at 300 fb-1

8

Higgs-Portal Interpretation

Hideki Okawa

• The limits on BR(H→inv) could be mapped to bounds on the coupling of Higgs-dark matter (DM) & DM-nucleon cross section for Higgs-portal DM models

• The Higgs-portal is a particular type of DM models, where DM interacts through the couplings to Higgs.

DM-nucleon scattering in Higgs-portal DM Model100

h

!

!

"h!!

(a)

h

N

!

N

!

"h!!

fN

(b)

Figure 65: Feynman diagrams for the decay of the Higgs boson into dark matter particles (a) and scat-

tering of dark matter particles off of a nucleon with the exchange of a Higgs boson (b). The Higgs-dark

matter interaction vertex has a coupling constant of !h"" . In the scattering diagram the Higgs-nucleon

coupling strength is parameterized with a form factor, fN .

#Scalar"N =! 2 Scalarh""

16$m4h

m4N f2N

!

m" +mN"2

(25)

#Vector"N =! 2 Vectorh""

16$m4h

m4N f2N

!

m" +mN"2

(26)

#Majorana"N =! 2 Majoranah""

4$%2m4h

m2"m4N f2N

!

m" +mN"2

(27)

The cross section has an additional dependence on the nucleon mass, mN and the form factor, fN1474

which quantifies the coupling strength between the Higgs boson and the Nucleon. This form factor is de-1475

termined using lattice calculations and suffers from large theoretical uncertainties [66]. These theoretical1476

uncertainties will not be included in the comparison plots.1477

Limits on both !h"" vs m" and #"N vs m" will be calculated from the invisible branching ratio1478

limits shown and will be compared to the limits from direct detection experiments. In calculating the1479

limits all variables in Equations 22- 27 are constants except for m" , !h"" , and #"N . The inputs used1480

for the remaining variables are given in Table 46. Limits on !h"" vs m" are shown in Figure 66 for1481

the scalar (66(a)), vector (66(b)), and majorana (66(b)) hypotheses. All direct detection results incur1482

a large uncertainty from the Higgs-Nucleon form factor uncertainty. Figure 67 shows limits on #"N1483

vs. m" . Direct detection results are published in this format and need no further interpretation. The1484

invisible branching ratio limits are shown for the scalar, vector, and majorana fermion hypothesis as1485

three curves. The hashed bands on the invisible branching ratio limits show the uncertainty resulting1486

from the systematic variation of fN .1487

It is evident from Figures 66 and 67 that the invisible branching fraction limits are complimentary1488

to the direct detection limits. Direct detection experiments provide the strongest limits at high mass, but1489

they loose all sensitivity below about 10 GeV. The invisible branching fraction limits are sensitive only1490

below mh/2 and provide exclusion below 10 GeV where the direct detection results do not reach. The1491

limits from the, scalar, vector, and fermion dark matter species depend differently on the dark matter1492

mass, but all exclude a large range of the coupling strength at low mass. Therefore, within the Higgs1493

portal model – which makes a generic assumption to test the higgs-dark matter coupling – the coupling1494

between dark matter and the higgs boson is strongly limited across a large range of dark matter mass.1495

Higgs decaying to DM

Our analysis Direct DM detection

experiments (XENON,

DAMA, etc.)


Hideki Okawa

Mapping & DM-types

�N�⇔ ⇔Higgs invisible decay Higgs-DM coupling DM-nucleon xsec

99

16 Dark Matter Interpretation of Branching Ratio Limits1437

One possible interpretation for an enhanced branching fraction to invisible particles is that the Higgs1438

boson decays to the dark matter particles that are expected to comprise approximately 24% of the energy1439

density of the universe [52, 53]. From cosmological observations a well motivated description of dark1440

matter is that it is weakly interacting and massive (WIMP). If the Higgs boson does decay to the dark1441

matter particle, then by virture of its interaction with the Higgs it would satisfy the WIMP hypothesis.1442

Many experiments have searched for dark matter by observing atoms recoiling from possible scatters of1443

dark matter particles. Direct detection experiments are sensitive to both the mass of the dark matter parti-1444

cle and its interaction cross section with nucleons in the atom and results are presented as a limit on these1445

paramters. Exclusion limits have been provided by a number of experiments including XENON [54, 55],1446

CDMSII [56], EDELWEISS [57, 58], ZEPLIN-III [59], COUPP [60], and SIMPLE [61]. Some experi-1447

ments have reported an observation of a dark matter signal, including CRESST [62], DAMA [63], and1448

CoGeNT [64]. The most recent observation from the CDMS collaboration [65] provides compelling1449

evidence an 8.6 GeV dark matter particle. Not all of the observations are consistent with each other and1450

some results are disputed by the community. Direct detection experiments make no a priori assumption1451

about the mechanism by which dark matter particles interact with Standard Model particles, but it is1452

possible that the interaction is through the exchange of a Higgs boson. If dark matter couples to the Stan-1453

dard Model through the Higgs boson and the mass of the particle is less than half the Higgs mass then1454

decays to the dark matter particle will enhance the invisible branching fraction. Under the assumption1455

that dark matter couples to the Standard Model only through the Higgs boson we aim to place limits1456

complimentary to the direct detection results on the mass and interaction cross section of the dark matter1457

particle.1458

Higgs Portal models [66, 67, 68] make a simple, ad-hoc extension to the Standard Model by intro-1459

ducing a new particle that couples to only the Higgs boson. The interaction strength is introduced with1460

a coupling constant, !h"" . Within this model the scattering and decay process can be compared by ex-1461

pressing the limits in terms of this coupling constant. Figure 65 shows feynman diagrams for both the1462

decay and scattering processes where !h"" appears in both diagrams. Using the feynman rules for these1463

diagrams the Higgs partial width and scattering cross section are determined in terms of !h"" . The Higgs1464

partial width for the decay to dark matter particles for the scalar, vector, and fermion cases is given in1465

Equations 22, 23, and 24 respectively.1466

#Scalar(h! "") =! 2 Scalarh"" v2

64$mh

!

1""

2m"

mh

#2$1/2

(22)

#Vector(h! "") =! 2 Vectorh"" v2

256$m4"mh

%

m4h"4m2"m2h+12m4"&

!

1""

2m"

mh

#2$1/2

(23)

#Majorana(h! "") =! 2 Majoranah"" v2mh

32$%2

!

1""

2m"

mh

#2$3/2

(24)

The partial width is a function of only the Higgs boson mass, the dark matter mass, the vacuum1467

expectation value, and the coupling constant. Note the introduction of a cutoff scale, % in the fermionic1468

case. In this case the Higgs interaction operator has dimension five and is non-renormalizable. A cutoff1469

scale is added that assumes the presence of new physics at a higher energy scale which would produce a1470

renormalizable theory. This model does not purport to be a complete model, so the addition of this cutoff1471

scale does not invaidate the model. For the scattering process the dark matter-nucleon cross section is1472

given for the for the scalar, vector, and fermion cases in Equations 25, 26, and 27 respectively.1473

100

h

!

!

"h!!

(a)

h

N

!

N

!

"h!!

fN

(b)

Figure 65: Feynman diagrams for the decay of the Higgs boson into dark matter particles (a) and scat-

tering of dark matter particles off of a nucleon with the exchange of a Higgs boson (b). The Higgs-dark

matter interaction vertex has a coupling constant of !h"" . In the scattering diagram the Higgs-nucleon

coupling strength is parameterized with a form factor, fN .

#Scalar"N =! 2 Scalarh""

16$m4h

m4N f2N

!

m" +mN"2

(25)

#Vector"N =! 2 Vectorh""

16$m4h

m4N f2N

!

m" +mN"2

(26)

#Majorana"N =! 2 Majoranah""

4$%2m4h

m2"m4N f2N

!

m" +mN"2

(27)

The cross section has an additional dependence on the nucleon mass, mN and the form factor, fN1474

which quantifies the coupling strength between the Higgs boson and the Nucleon. This form factor is de-1475

termined using lattice calculations and suffers from large theoretical uncertainties [66]. These theoretical1476

uncertainties will not be included in the comparison plots.1477

Limits on both !h"" vs m" and #"N vs m" will be calculated from the invisible branching ratio1478

limits shown and will be compared to the limits from direct detection experiments. In calculating the1479

limits all variables in Equations 22- 27 are constants except for m" , !h"" , and #"N . The inputs used1480

for the remaining variables are given in Table 46. Limits on !h"" vs m" are shown in Figure 66 for1481

the scalar (66(a)), vector (66(b)), and majorana (66(b)) hypotheses. All direct detection results incur1482

a large uncertainty from the Higgs-Nucleon form factor uncertainty. Figure 67 shows limits on #"N1483

vs. m" . Direct detection results are published in this format and need no further interpretation. The1484

invisible branching ratio limits are shown for the scalar, vector, and majorana fermion hypothesis as1485

three curves. The hashed bands on the invisible branching ratio limits show the uncertainty resulting1486

from the systematic variation of fN .1487

It is evident from Figures 66 and 67 that the invisible branching fraction limits are complimentary1488

to the direct detection limits. Direct detection experiments provide the strongest limits at high mass, but1489

they loose all sensitivity below about 10 GeV. The invisible branching fraction limits are sensitive only1490

below mh/2 and provide exclusion below 10 GeV where the direct detection results do not reach. The1491

limits from the, scalar, vector, and fermion dark matter species depend differently on the dark matter1492

mass, but all exclude a large range of the coupling strength at low mass. Therefore, within the Higgs1493

portal model – which makes a generic assumption to test the higgs-dark matter coupling – the coupling1494

between dark matter and the higgs boson is strongly limited across a large range of dark matter mass.1495

�(h� ��) �2h��

BR(h� ��) =�(h� ��)

�(h� ��) + �(h� SM)

We consider three DM types: scalar, vector, majorana fermion

9Snowmass: Seattle Energy Frontier Workshop, July 2, 2013

10

BR(H→inv) to Higgs-Portal

Hideki Okawa

• Mapped BR(H→inv)=10% line (as a benchmark) to Higgs-portal DM interpretation

• Very good sensitivity in mχ<mH/2 region

• Uncertainty from the nucleon form factor is shown (left plot)

Mass [GeV]1 10 210 310

]2D

M-N

ucle

on c

ross

-sec

iton

[cm

-5110-5010-4910-4810-4710-4610-4510-4410-4310-4210-4110-4010-3910-3810

XENON 100

XENON 10

XENON 1TDAMA/LIBRA

CRESST

mCDMS 2

mCDMS 1CoGeNT

inv.)=10%, scalar DMABR(H

inv.)=10%, vector DMABR(H

inv.)=10%, fermion DMABR(H

If BR(H→inv) < 10%

DM-nucleon cross section

Mass [GeV]1 10 210 310

) rrhh

DM

-Hig

gs c

oupl

ing

(

-310

-210

-110

1

10

210

XENON 100XENON 10XENON 1TDAMA/LIBRACRESST

mCDMS 1mCDMS 2

CoGeNT inv.)=10%ABR(H

Scalar

DM-Higgs Coupling

Mass [GeV]1 10 210 310

) rrhh

DM

-Hig

gs c

oupl

ing

(

-610

-510

-410

-310

-210

-1101

10

210

310

410XENON 100XENON 10XENON 1TDAMA/LIBRACRESST

mCDMS 1mCDMS 2


Vector

Mass [GeV]1 10 210 310

)R

/ rrhh

DM

-Hig

gs c

oupl

ing

(

-510

-410

-310

-210

-110

1XENON 100XENON 10XENON 1TDAMA/LIBRACRESST

mCDMS 1mCDMS 2


Majorana fermion


11

Summary & Plans

Hideki Okawa

• Showed preliminary studies on the prospects of LHC for ZH→ll+invisible channel.

• Considered benchmark luminosities & μ-values proposed by the Snowmass committee.

• As this channel significantly relies on the performance of Missing ET, improving the pileup suppression in the Missing ET calculation would have quite an impact on the signal sensitivity.

• As long as ETmiss is under control, the main background is ZZ. The systematics of ZZ will be the key component for the signal sensitivity.

• Detailed investigations are ongoing, and expected limits for the Snowmass scenarios are to be provided.


backups

Moriond Results

Hideki Okawa 13

[GeV]missTE

0 50 100 150 200 250 300 350 400 450

Even

ts /

30 G

eV

0

10

20

30

40

50ATLAS =8 TeVsPreliminary -1 L = 13.0 fb0Data

ZTop

WZWW

ZZ Other BG=125 GeV)

HSignal (SM ZH, m

[GeV]missTE

0 50 100 150 200 250 300 350 400 450

Even

ts /

30 G

eV

5

10

15

20

25ATLAS =7 TeVsPreliminary -1 L = 4.7 fb0Data

ZTop

WZWW

ZZ Other BG=125 GeV)

HSignal (SM ZH, m

• Consistent with the SM predictions.

• Limits are set on the two scenarios as mentioned in the next slide.

From Moriond CONF note


Hideki Okawa

Moriond Results

14

• ZZ, WZ are dominated by the jet systematics & theory uncertainty

• Z uncertainty comes from both the statistical and systematical uncertainty.

From Moriond CONF note


Documents

LHC Potential for ZH ll + invisible Search€¦ · • The ﬁrst direct search for the invisible Higgs at the LHC. BR(H→inv)