search for dark matter in events with a single boson and missing...

An Alpine LHC Physics Summit (ALPS 2017), April 20, 2017

Hideki Okawa for the ATLAS Collaboration University of Tsukuba, Division of Physics & CiRfSE

Search for Dark Matter in Events with a Single Boson and

Missing Transverse Momentum using the ATLAS Detector

©NASA

ALPS2017, April 20, 2017Hideki Okawa

Dark Matter Searches

2

DM Mass [GeV]1 10 210 310

]2 (D

M-p

roto

n) [c

mSD

σ

46−10

45−10

44−10

43−10

42−10

41−10

40−10

39−10

38−10

37−10

36−10

35−10 DM Simplified Model Exclusions Preliminary March 2017ATLAS

= 1DM

= 0, gl

= 0.25, gqgAxial-vector mediator, Dirac DM

ATLAS limits at 95% CL, direct detection limits at 90% CL

Dijet TLA

ATLAS-CONF-2016-030-1 = 13 TeV, 3.4 fbs

Dijet TLA

Dijet 8 TeV

Phys. Rev. D. 91 052007 (2015)-1 = 8 TeV, 20.3 fbs

Dijet 8 TeV

Dijet

arXiv:1703.09127 [hep-ex]-1 = 13 TeV, 37.0 fbs

Dijetγ+miss

TE

CERN-EP-2017-044

-1 = 13 TeV, 36.4 fbs

γ+missTE

8F3PICO-60 C

arXiv:1702.07666v1 [astro-ph.CO]8F3PICO-60 C

WIMP mass [GeV]1 10 210 310

]2W

IMP-

nucl

eon

cros

s se

ctio

n [c

m

57−10

55−10

53−10

51−10

49−10

47−10

45−10

43−10

41−10

39−10

DAMA/LIBRA (99.7% CL)CRESST II (95% CL)CDMS SI (95% CL)CoGeNT (99% CL)CRESST II (90% CL)SuperCDMS (90% CL)XENON100 (90% CL)LUX (90% CL)

Scalar WIMPMajorana WIMPVector WIMP

ATLAS

Higgs portal model:ATLAS 90% CL in

-1 = 7 TeV, 4.5-4.7 fbs-1 = 8 TeV, 20.3 fbs

Vis. & inv. Higgs boson decay channels]inv, BRγZκ, γκ, gκ, µκ, τκ, bκ, tκ, Zκ, Wκ[

<0.22 at 90% CLinv

assumption: BRW,ZκNo

JHEP11(2015)206

• Dark matter (DM) has been extensively searched by direct, indirect detection and previous collider experiments (e.g. PICO-60 C3F8, LUX, PandaX-II, Xenon-100, Tevatron).

• LHC offers complementary sensitivities to some specific scenarios (e.g. axial-vector DM mediator, Higgs-portal).

(talks by Graciela Beatriz Gelmini, Florian Reindl)

(talks by Steven Lowette, Patrick Fox)

http://link.springer.com/article/10.1007/JHEP11%282015%29206


ALPS2017, April 20, 2017Hideki Okawa 3

q

qZ

0 h

A

5. the coupling constant between the Z 0, Higgs boson and pseudoscalar fields, gZ 0.

2.2 Heavy scalar model

t

t

tg

g

H H

h

Figure 2: Production of a heavy scalar H via gluon fusion (left), and decay of the H into a Higgs boson (denoted ash) and a pair of dark matter candidates via an eective coupling (right).

The heavy scalar model [10] introduces a heavy scalar H 1 in the mass range of 2mh < mH < 2mtop, whichis produced primarily via gluon fusion (ggF) as shown in Fig. 2 (left). The upper bound on mH is to avoidlarge tt branching of the H . The DM mass m is taken to be roughly half of the SM Higgs boson mass toensure on shell decay of H ! h , and suppress invisible decay modes of h. It also satisfies constraintsfrom direct detection experiments [11]. The heavy scalar is allowed to decay into a Higgs boson andDM candidates, H ! h , where is a DM candidate with spin-0, as shown in Fig. 2 (right). Theseinteractions can be expressed by the Lagrangian:

LQ = 12

Hh Hh 14

HHhhHHhh 1

4

hh hh 14HH HH , (1)

the first term of which contains an eective quartic coupling between h, H , and , relevant for thisanalysis. The larger branching ratio (BR) of the three-body decay of H to the SM Higgs boson and DMparticles compared to two-body decays can be obtained by assuming that the DM candidates originatefrom the decay of some real scalar S. The masses mH and m , as well as the branching ratios of the threedecay modes of H , are free parameters of the model.

3 The ATLAS Detector

The ATLAS detector [12] is a multi-purpose particle physics detector with approximately forward-backward symmetric cylindrical geometry2. The inner tracking detector (ID) covers | | < 2.5 andconsists of a silicon pixel detector, a silicon micro-strip detector, and a transition radiation tracker (TRT)which allow a precise reconstruction of charged particle trajectories and of decay vertices of long lifetimeparticles. This is especially important for the present analysis in the case of photon conversion vertexreconstruction. The tracking system includes the newly installed innermost pixel layer, the insertableB-Layer (IBL) [13]. The ID is surrounded by a thin superconducting solenoid providing a 2 T axialmagnetic field. A high-granularity lead/liquid-argon (LAr) sampling calorimeter measures the energy1 Throughout this note, a lowercase h refers to the observed Higgs boson, while an uppercase H refers to the heavy scalar in

this model.2 ATLAS uses a right-handed coordinate system with its origin at the nominal interaction point (IP) in the centre of the detector

and the z-axis along the beam pipe. The x-axis points from the IP to the centre of the LHC ring, and the y axis pointsupward. Cylindrical coordinates (r, ) are used in the transverse plane, being the azimuthal angle around the beam pipe.The pseudorapidity is defined in terms of the polar angle as = ln tan(/2).

4

Mediator Mass [TeV]0 0.5 1 1.5 2 2.5 3

DM M

ass

[TeV

]

0.2

0.4

0.6

0.8

1

1.2DM Simplified Model Exclusions Preliminary March 2017ATLAS

= 1DM

= 0, gl

= 0.25, gq

gAxial-vector mediator, Dirac DM

All limits at 95% CL

Pertu

rbat

ive U

nitar

ity

Dije

t

arXiv:1703.09127 [hep-ex]

-1 = 13 TeV, 37.0 fbsDijet

Dije

t 8 T

eVPhys. Rev. D. 91 052007 (2015)

-1 = 8 TeV, 20.3 fbsDijet 8 TeV

Dije

t TLA

ATLAS-CONF-2016-030

-1 = 13 TeV, 3.4 fbsDijet TLA

Dije

t + IS

R

ATLAS-CONF-2016-070

-1 = 13 TeV, 15.5 fbsDijet + ISR

γ+missTE

CERN-EP-2017-044

-1 = 13 TeV, 36.4 fbsγ+miss

TE

+jetmissTE JHEP 06 (2016) 059

-1 = 13 TeV, 3.2 fbs+jetmiss

TE

DM Mas

s = M

ediat

or Mas

s

×2

= 0.1

22hcΩ

Therm

al Relic

• Among the mono-X searches, the monojet channel is generally the most sensitive channels, followed by the monophoton (talk by William Kalderon).

• Nevertheless, mono-W/Z and mono-H channels are important when the DM or mediator only interacts with the vector bosons or Higgs boson.

• In these searches, identification/reconstruction of the bosons & missing transverse momentum (ETmiss): indirect measurement of pT(𝝌𝝌) are the key.

Mono-X Searches

5. the coupling constant between the Z 0, Higgs boson and pseudoscalar fields, gZ 0.

2.2 Heavy scalar model

t

t

tg

g

H H

h

Figure 2: Production of a heavy scalar H via gluon fusion (left), and decay of the H into a Higgs boson (denoted ash) and a pair of dark matter candidates via an eective coupling (right).

The heavy scalar model [10] introduces a heavy scalar H 1 in the mass range of 2mh < mH < 2mtop, whichis produced primarily via gluon fusion (ggF) as shown in Fig. 2 (left). The upper bound on mH is to avoidlarge tt branching of the H . The DM mass m is taken to be roughly half of the SM Higgs boson mass toensure on shell decay of H ! h , and suppress invisible decay modes of h. It also satisfies constraintsfrom direct detection experiments [11]. The heavy scalar is allowed to decay into a Higgs boson andDM candidates, H ! h , where is a DM candidate with spin-0, as shown in Fig. 2 (right). Theseinteractions can be expressed by the Lagrangian:

LQ = 12

Hh Hh 14

HHhhHHhh 1

4

hh hh 14HH HH , (1)

the first term of which contains an eective quartic coupling between h, H , and , relevant for thisanalysis. The larger branching ratio (BR) of the three-body decay of H to the SM Higgs boson and DMparticles compared to two-body decays can be obtained by assuming that the DM candidates originatefrom the decay of some real scalar S. The masses mH and m , as well as the branching ratios of the threedecay modes of H , are free parameters of the model.

3 The ATLAS Detector

The ATLAS detector [12] is a multi-purpose particle physics detector with approximately forward-backward symmetric cylindrical geometry2. The inner tracking detector (ID) covers | | < 2.5 andconsists of a silicon pixel detector, a silicon micro-strip detector, and a transition radiation tracker (TRT)which allow a precise reconstruction of charged particle trajectories and of decay vertices of long lifetimeparticles. This is especially important for the present analysis in the case of photon conversion vertexreconstruction. The tracking system includes the newly installed innermost pixel layer, the insertableB-Layer (IBL) [13]. The ID is surrounded by a thin superconducting solenoid providing a 2 T axialmagnetic field. A high-granularity lead/liquid-argon (LAr) sampling calorimeter measures the energy1 Throughout this note, a lowercase h refers to the observed Higgs boson, while an uppercase H refers to the heavy scalar in

this model.2 ATLAS uses a right-handed coordinate system with its origin at the nominal interaction point (IP) in the centre of the detector

and the z-axis along the beam pipe. The x-axis points from the IP to the centre of the LHC ring, and the y axis pointsupward. Cylindrical coordinates (r, ) are used in the transverse plane, being the azimuthal angle around the beam pipe.The pseudorapidity is defined in terms of the polar angle as = ln tan(/2).

4

q

q/q′

W/Z

W/Z

χ

χ



Missing Transverse Momentum

4

CA#Lee,#P#Loch# # #Topical#Meeting#on#ETmiss#in#Physics#Analyses# ############# ####23#June#2015

• Object#terms#are#calculated#from#the#fully(reconstructed(&(calibrated#hard#objects#in#the#event,#in#the#order#shown#in#the#diagram#and#equation#– The#order#is#vital#to#avoid#overlap,#and#establishes#a#priority#of#the#terms#!

• Tracks#not#associated#to#any#hard#object,#and#associated#to#the#primary#vertex,#form#the#TrackFbased(Soft(Term#(TST)##– This(is(currently(the(recommended(soft(term(for(use(in(run(2(!

• Calorimeter#topoclusters#not#associated#to#any#hard#object,#and#calibrated#using#LCW,#form#the#Calorimeter(Soft(Term((CST)#!

• The#MET#is#the#vectorial#sum#of#the#hard#and#soft#terms:#!!

• TrackMET:#!#and#μ#objects#+#jet#tracks#+#TST

4

MET reconstruction in run 2:

Difference with MET in run 1: NO default hard object selections (except jets)

©ATLAS MET Group

• Missing transverse momentum (ETmiss) is the pT imbalance of reconstructed physics objects.

Emissx(y) = pe

x(y) px(y) p

had,visx(y) pjets

x(y) pµx(y) psoft

x(y)

• It is indirect measurement of weakly interacting particle (neutrinos, dark matter) momenta, so very important for dark matter searches.

PVN0 5 10 15 20 25 30

Res

olut

ion

[GeV

]ym

iss

,Exm

iss

E

05

101520253035404550

ATLAS Simulation Preliminary

= 13 TeVs 25nsµµ →Z

all jets

missTCST EmissTTST E

missTTrack E

[GeV]missTE

Even

ts/3

0 G

eV

2−101−101

10210310410510610710810910

101011101210

[GeV]missTE

0 100 200 300 400 500 600 700 800 900Data

/Pre

d.

00.5

11.5

2

DataZZWZWW

)+jetsµµZ(ee)/Z(TopFake LeptonOther BkgsStat.+Sys.

ATLAS Preliminary-1=13 TeV, 13.3 fbsµµ ee+

ATLAS-CONF-2016-056ATL-PHYS-PUB-2015-023

(Run 1)(Run 2)

TST ETmiss

CST ETmiss: Algorithm using Calorimeter Soft Term

TST ETmiss: Algorithm using Track Soft Term

Track ETmiss: Purely reconstructed from tracks

https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2016-056/

http://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PUBNOTES/ATL-PHYS-PUB-2015-023


Boosted Boson Tagging

5

V Vlow-pT V (2 jets)

[GeV]TWp

0 100 200 300 400 500 600 700 800

)q R

(q,

∆

0

0.5

1

1.5

2

2.5

3

3.5

4

020406080100120140160180200

Wb→, t t t→Pythia Z' ATLAS Simulation

JHEP09 (2013) 076

high-pT V (1 jet)

Trimming

Jet mass [GeV]0 20 40 60 80 100 120 140 160 180 200

Nor

mal

ised

to u

nity

00.020.040.060.08

0.10.120.140.160.180.2

0.22

< 350 GeVTruthT

200 < pW-jetsZ-jetsMultijets

< 2000 GeVTruthT

1500 < pW-jetsZ-jetsMultijets

ATLAS Simulation Preliminary R = 1.0 jetstanti-k

= 0.2)sub = 5%, Rcut

Trimmed (f | < 2.0Truthη |

= 13 TeVs

Mass [GeV]0 50 100 150 200 250

Arbi

trary

Uni

ts0

0.050.1

0.150.2

0.250.3

0.350.4

0.450.5

Higgs-jet

Multi-jet

Hadronic top jet

R=1.0 jetstanti-k=0.2)sub=0.05, R

cutTrimmed (f

| < 2.0detη < 500 GeV, |

T350 < p

Double b-tagged @ 70% WP

ATLAS Simulation Preliminary

ATL-PHYS-PUB-2015-035ATL-PHYS-PUB-2015-033

• Anti-kt large-R jet (R=1.0) is used. Trimming is performed to remove pileup contributions.

• Cut on D2 (energy correlation ratio; offers separation for 1- & 2-prong decays) is applied to reduce the multijet BG.

• Jets with mass consistent with mV or mH are selected.

• b-tagging using small-R track jets (R=0.2) is further considered to tag the Higgs bosons.

R~2MW/pTW

https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PUBNOTES/ATL-PHYS-PUB-2015-035/


Mono-W/Z(qq)

6

q

q/q′

W/Z

W/Z

χ

χ q

q/q′

med

W/Z χ

χ

Phys. Lett. B 763 (2016) 251

• Hadronic final states have advantage of large yields.

• BR(W→qq) ~ 3BR(W→eve, μνμ)

• BR(Z→qq) ~ 10BR(Z→ee, μμ)

• No lepton.

• ETmiss > 250 GeV & Track ETmiss > 30 GeV.

• dϕ(ETmiss, Track ETmiss) < 𝜋/2.

• Boosted boson tagging:

• Large-R jet pT>200 GeV, |η|<2.0. Trimmed.

• D2 (2-prong-ness) & jet mass consistent w/ W or Z. → pT-dependent D2 cut w/ 50% tagging efficiency.

2015 Data

(3.2 fb-1 )

Even

ts /

0.2

200

400

600

800

1000

1200

1400

1600Data 2015Z+jetsW+jets + Single Toptt

DibosonUncertainty (stat. + syst.)

=1000,1995GeVmed,mχ, m4vec. W*10

ATLAS = 13 TeVs

-1Ldt = 3.2 fb∫Signal region

2D0 0.5 1 1.5 2 2.5 3D

ata/

MC

0.51

1.5

http://www.sciencedirect.com/science/article/pii/S037026931630613X


Mono-W/Z(qq)

7

• Main BGs: Z(vv)+jets, W(ℓv)+jets, Top, dibosons.

• Normalization is taken from control regions for Z(vv)+jets, W(ℓv)+jets & Top.

• Z control region: 2 muons, 66<mμμ<116 GeV.

• W (Top) control region: 1 muon, 0 b-jet (≥1 b-jets).

• ET no μmiss: muon contributions removed from ETmiss.

• Fully MC for shape & dibosons/single-top.

2015 Data

(3.2 fb-1 )

[GeV]missT no muE

200 400 600 800 1000 1200 1400

Even

ts /

GeV

-210

-110

1

10

210Data 2015Z+jetsW+jetsSingle topttDibosonUncertaintyPre-fit background

ATLAS -1Ldt = 3.2 fb∫ = 13 TeV s

Z(ll)+jets control region

[GeV]missT no muE

200 400 600 800 1000 1200 1400Dat

a/Bk

g

0.80.9

11.11.2

[GeV]missT no muE

200 400 600 800 1000 1200 1400

Even

ts /

GeV

-210

-110

1

10

210

310 Data 2015Z+jetsW+jetsSingle topttDibosonUncertaintyPre-fit background


control region tt

[GeV]missT no muE

200 400 600 800 1000 1200 1400Dat

a/Bk

g

0.80.9

11.11.2

[GeV]missT no muE

200 400 600 800 1000 1200 1400

Even

ts /

GeV

-210

-110

1

10

210

310Data 2015Z+jetsW+jetsSingle topttDibosonUncertaintyPre-fit background


W+jets control region

[GeV]missT no muE

200 400 600 800 1000 1200 1400Dat

a/Bk

g

0.80.9

11.11.2


Mono-W/Z(qq)

8

[GeV]χ m0 200 400 600 800 1000

[GeV

]*

95%

C.L

. low

er li

mit

on M

400

500

600

700

800

900

1000

ObservedExpected

σ 1±

σ 2±

ATLAS-1 L=3.2 fb∫ = 13 TeVs

EFTχχVV

[GeV]medm200 400 600 800 1000 1200 1400 1600 1800 2000

[GeV

]χ

m

100

200

300

400

500

600

700

800

900

1000 µ95

% C

.L. u

pper

lim

it on

1

10

210ATLAS

-1 L=3.2 fb∫ = 13 TeVs+W/Z: vector modelmiss

TE

=1DM

=0.25, gSM

g

[GeV]missTE

400 600 800 1000 1200 1400

Even

ts /

GeV

-310

-210

-110

1

10

210

Data 2015+W/Z: vector modelmiss

TE=10 TeVmed=10 GeV, mDMm

Z+jetsW+jetsSingle topttDibosonUncertaintyPre-fit background


signal region

[GeV]missTE

400 600 800 1000 1200 1400Dat

a/Bk

g

00.5

11.5

2

• No deviation from the SM prediction.

• Interpreted with an EFT and simplified vector mediator model.

2015 Data

(3.2 fb-1 )


Mono-Z(ℓℓ)

9

[GeV]missTE

Even

ts/1

0 G

eV

3−10

2−10

1−10

1

10

210

310

410

[GeV]missTE

210 310

Data

/Pre

d.

00.5

11.5

2

DataZZWZOther BkgsZHinv.

= 10 GeVmed

= 1 GeV, mχm = 300 GeV

med = 50 GeV, mχm

non-resonant-ll)+jetsµµZ(ee)/Z(

Fake LeptonStat.+Sys.

ATLAS Preliminary-1=13 TeV, 13.3 fbs

LMSR µµee+

[GeV]medm0 100 200 300 400 500 600 700

[GeV

]χ

m

0

20

4060

80

100

120140

160

180

χ

= 2 m

med

m

-1=13 TeV, 13.3 fbs = 1.0χ

= 0.25, gqVector, Dirac, g

PreliminaryATLAS )σ1 ±Expected limit (

Observed limit

• This channel is also sensitive to the Higgs-portal model.

• BR(H→inv) < 0.98 [obs], 0.65 [exp] at 95% CL. Not yet surpassing the Run-1 constraint (< 0.75 [obs], 0.62 [exp]).

• Clean channel, but lacks in statistics.

• Z+jets BG is highly suppressed by various kinematic selections (ETmiss > 90 GeV, etc.). The irreducible ZZ BG is the most dominant.

• No significant excess from the SM prediction.

2015 + 2016

(13.3 fb-1 )


Mono-H Searches

10

q

q

Z0

Z0

h

q

qZ

0 h

A

• Higgs-strahlung from initial-state partons is suppressed by the Yukawa coupling.

• Mono-H searches are direct probes for the DM interactions.

• Effective Field Theory (Run-1) & various simplified models (Run-2) have been considered in ATLAS.

H

g

g

h

Vector Mediator Model

gq g𝝌gZ′

Z′-2HDM Model Heavy Scalar Model


Mono-H(𝛾𝛾)

11

ATLAS-CONF-2017-024

]GeV [TmissES

0 5 10 15 20

Dat

a / M

C

012

GeV

Even

ts /

1−10

1

10

210

310

410

510

610

710 Data

= 200 GeV0Am = 1 TeV, Z'm = 100 GeV,χmZ'-2HDM, Dirac DM

= 200 GeVBZ'm = 1 GeV, χm, Dirac DM BZ'

= 275 GeVHm = 60 GeV,χmHeavy scalar, Scalar DM

γγV

γV

SM Higgs boson

+jetsγ

γγ

Stat. Unc.⊕Syst.

PreliminaryATLAS-1 = 13 TeV, 36.1 fbs

• Low BR, but very clean signature.

• Baseline selection is the same as the SM H→𝛾𝛾: • At least two photons with pT>25 GeV.

• ET𝛾1/m𝛾𝛾>0.35, ET𝛾2/m𝛾𝛾>0.25

• 105 < m𝛾𝛾 < 160 GeV

• 5 event categories based on SETmiss, pT𝛾𝛾, pThard, # of leptons, |zPVhard-zPV𝛾𝛾|

• First category for vector-mediator & Z′-2HDM.

• All categories for heavy scalar models.

Category Requirements

Mono-Higgs SEmissT

> 7pGeV, pT > 90 GeV, lepton veto

High-EmissT SEmiss

T> 5.5

pGeV, |zhardPV zPV | < 0.1 mm

Intermediate-EmissT SEmiss

T> 4

pGeV, phardT > 40 GeV , |zhardPV zPV | < 0.1 mm

Di↵erent-Vertex SEmissT

> 4pGeV, phardT > 40 GeV , |zhardPV zPV | > 0.1 mm

Rest pT > 15 GeV

• the EmissT significance, SEmiss

T= Emiss

T /pP

ET. The sum in the denominator is the total transverseenergy deposited in the calorimeters or in the muon detectors in the event3;

• the diphoton transverse momentum, pT ;

• phardT ;

• the number of leptons in the event;

• the distance between the diphoton vertex and the highest p2T vertex in the z direction: |zhard

PV zPV |.The resulting categorization scheme is shown in Table 1. Each category excludes events that are in theupper categories. The Z 0B and Z 0-2HDM signal samples have been used to optimize the definition ofthe Mono-Higgs category, which provides most of the sensitivity to those two models. The other fourcategories, which provide extra sensitivity to heavy scalar boson events with softer Emiss

T , have beenoptimized using simulated heavy scalar boson samples to cover the dierent kinematic regimes of theheavy scalar model.

Table 1: Optimized criteria used in the categorization. Each category excludes events that are in the upper categories.

Category Requirements

Mono-Higgs SEmissT> 7p

GeV, pT > 90 GeV, lepton vetoHigh-Emiss

T SEmissT> 5.5

pGeV, |zhard

PV zPV | < 0.1 mmIntermediate-Emiss

T SEmissT> 4p

GeV, phardT > 40 GeV, |zhard

PV zPV | < 0.1 mmDierent-Vertex SEmiss

T> 4p

GeV, phardT > 40 GeV, |zhard

PV zPV | > 0.1 mmRest pT > 15 GeV

Figure 2 shows the distributions of SEmissT

, phardT and pT after the selection of diphoton candidates in

120 GeV < m < 130 GeV. Expected signal distributions are shown for a Z 0B signal with mZ0B= 200 GeV

and Dirac fermion DM m = 1 GeV, a Z 0-2HDM signal with mZ0 = 1000 GeV, mA0 = 200 GeV andDirac fermion DM m = 100 GeV, and a heavy scalar model with mH = 275 GeV and scalar DM m = 60GeV. These overlaid signal points chosen are representative of the model kinematics.

The non-resonant and +jet samples account for 79% and 19% of the non-resonant background inthe signal region. The shape of the +jet sample in the distributions is obtained by extrapolation froma data control sample where one photon fails the photon identification criteria. This control sampleis mainly dominated by +jet contributions but there is a non-negligible contamination coming fromnon-resonant , V, and V. This contamination is estimated by applying the same inverted photonidentification criteria. Then, it is subtracted from the control sample so that a reliable +jet shape isobtained. Considering that the ratio between +jet and sample is constant in the control region andthe signal region, +jet is estimated in the signal region by applying +jet/ weights to sample. Theslight discrepancies observed in the distributions of SEmiss

T, pT and phard

T in Fig. 2 do not aect the result.

3 The total transverse energy is calculated from the scalar sum of the transverse momenta of the calibrated reconstructed objects(photons, electrons, muons and jets) used in the Emiss

T calculation described in Sec. 4, as well as the tracks not associatedwith any of the hard objects.

8

pThard = | |

Eur. Phys. J. C (2017) 77:241 Page 25 of 46 241

)PV

Number of Reconstructed Vertices (N0 5 10 15 20 25 30 35 40

RM

S [G

eV]

mis

sy

,Em

iss

xE

0

5

10

15

20

25

30

35

> 20 GeVTJet P > 25 GeVTJet P > 30 GeVTJet P > 35 GeVTJet P

Inclusive

ATLAS Simulation = 8 TeV, Jet JVF > 0.25s

µµ→Z

[GeV]ZT

p0 10 20 30 40 50 60 70 80 90 100

[GeV

]⟩

ZA⋅m

iss

TE⟨TS

T

25−

20−

15−

10−

5−

0

5

10


ATLAS Simulation = 8 TeV, JVF > 0.25s

Inclusiveµµ→Z

(a) (b)

Fig. 19 The a TST EmissT resolution as a function of the number of

reconstructed vertices per bunch crossing (NPV) and the b TST EmissT in

the direction of the pZT are shown for different jet-pT thresholds using

the Z → µµ simulation. JVF> 0.25 is required for all jets with pT > 20GeV and |η| < 2.4

EmissT becomes strongly biased in the direction opposite to

the pZT . Therefore, the pT threshold of 20 GeV is preferred.

8 Systematic uncertainties of the soft term

The EmissT is reconstructed from the vector sum of sev-

eral terms corresponding to different types of contributionsfrom reconstructed physics objects, as defined in Eq. (2).The estimated uncertainties in the energy scale and momen-tum resolution for the electrons [14], muons [13], jets [44],τhad-vis [47], and photons [14] are propagated into the Emiss

T .This section describes the estimation of the systematic uncer-tainties for the Emiss

T soft term. These uncertainties take intoaccount the impact of the generator and underlying-eventmodelling used by the ATLAS Collaboration, as well aseffects from pileup.

The balance of the soft term with the calibrated physicsobjects is used to estimate the soft-term systematic uncer-tainties in Z → µµ events, which have very little Emiss,True

T .The transverse momenta of the calibrated physics objects,p hard

T , is defined as

p hardT =

!p e

T +!

p µT +

!p γ

T +!

p τT +

!p jet

T ,

(15)

which is the vector sum of the transverse momenta of thehigh-pT physics objects. It defines an axis (with unit vectorp hard

T ) in the transverse plane of the ATLAS detector alongwhich the Emiss

T soft term is expected to balance phardT in

Z → µµ events. This balance is sensitive to the differences incalibration and reconstruction of the Emiss,soft

T between dataand MC simulation and thus is sensitive to the uncertainty inthe soft term. This discussion is similar to the one in Sect. 6.2;

however, here the soft term is compared to the hard termrather than comparing the Emiss

T to the recoil of the Z .

8.1 Methodology for CST

Two sets of systematic uncertainties are considered for theCST. The same approach is used for the STVF and EJAFalgorithms to evaluate their soft-term systematic uncertain-ties. The first approach decomposes the systematic uncertain-ties into the longitudinal and transverse components along thedirection of p hard

T , whereas the second approach estimates theglobal scale and resolution uncertainties. While both meth-ods were recommended for analyses of the 8 TeV dataset, thefirst method, described in Sect. 8.1.1, gives smaller uncer-tainties. Therefore, the second method, which is discussed inSect. 8.1.2, is now treated as a cross-check.

Both methods consider a subset of Z → µµ events thatdo not have any jets with pT > 20 GeV and |η| < 4.5. Suchan event topology is optimal for estimation of the soft-termsystematic uncertainties because only the muons and the softterm contribute to the Emiss

T . In principle the methods arevalid in event topologies with any jet multiplicity, but theZ → µµ+ ≥1-jet events are more susceptible to jet-relatedsystematic uncertainties.

8.1.1 Evaluation of balance between the soft term and thehard term

The primary or “balance” method exploits the momentumbalance in the transverse plane between the soft and hardterms in Z → ℓℓ events, and the level of disagreementbetween data and simulation is assigned as a systematicuncertainty.

The E miss,softT is decomposed along the p hard

T direction.The direction orthogonal to p hard

T is referred to as the per-

123

Eur. Phys. J. C (2017) 77:241 Page 25 of 46 241

)PV

Number of Reconstructed Vertices (N0 5 10 15 20 25 30 35 40

RM

S [G

eV]

mis

sy

,Em

iss

xE

0

5

10

15

20

25

30

35


Inclusive

ATLAS Simulation = 8 TeV, Jet JVF > 0.25s

µµ→Z

[GeV]ZT

p0 10 20 30 40 50 60 70 80 90 100

[GeV

]⟩

ZA⋅m

iss

TE⟨TS

T

25−

20−

15−

10−

5−

0

5

10


ATLAS Simulation = 8 TeV, JVF > 0.25s

Inclusiveµµ→Z

(a) (b)

Fig. 19 The a TST EmissT resolution as a function of the number of

reconstructed vertices per bunch crossing (NPV) and the b TST EmissT in

the direction of the pZT are shown for different jet-pT thresholds using

the Z → µµ simulation. JVF> 0.25 is required for all jets with pT > 20GeV and |η| < 2.4

EmissT becomes strongly biased in the direction opposite to

the pZT . Therefore, the pT threshold of 20 GeV is preferred.

8 Systematic uncertainties of the soft term

The EmissT is reconstructed from the vector sum of sev-

eral terms corresponding to different types of contributionsfrom reconstructed physics objects, as defined in Eq. (2).The estimated uncertainties in the energy scale and momen-tum resolution for the electrons [14], muons [13], jets [44],τhad-vis [47], and photons [14] are propagated into the Emiss

T .This section describes the estimation of the systematic uncer-tainties for the Emiss

T soft term. These uncertainties take intoaccount the impact of the generator and underlying-eventmodelling used by the ATLAS Collaboration, as well aseffects from pileup.

The balance of the soft term with the calibrated physicsobjects is used to estimate the soft-term systematic uncer-tainties in Z → µµ events, which have very little Emiss,True

T .The transverse momenta of the calibrated physics objects,p hard

T , is defined as

p hardT =

!p e

T +!

p µT +

!p γ

T +!

p τT +

!p jet

T ,

(15)

which is the vector sum of the transverse momenta of thehigh-pT physics objects. It defines an axis (with unit vectorp hard

T ) in the transverse plane of the ATLAS detector alongwhich the Emiss

T soft term is expected to balance phardT in

Z → µµ events. This balance is sensitive to the differences incalibration and reconstruction of the Emiss,soft

T between dataand MC simulation and thus is sensitive to the uncertainty inthe soft term. This discussion is similar to the one in Sect. 6.2;

however, here the soft term is compared to the hard termrather than comparing the Emiss

T to the recoil of the Z .

8.1 Methodology for CST

Two sets of systematic uncertainties are considered for theCST. The same approach is used for the STVF and EJAFalgorithms to evaluate their soft-term systematic uncertain-ties. The first approach decomposes the systematic uncertain-ties into the longitudinal and transverse components along thedirection of p hard

T , whereas the second approach estimates theglobal scale and resolution uncertainties. While both meth-ods were recommended for analyses of the 8 TeV dataset, thefirst method, described in Sect. 8.1.1, gives smaller uncer-tainties. Therefore, the second method, which is discussed inSect. 8.1.2, is now treated as a cross-check.

Both methods consider a subset of Z → µµ events thatdo not have any jets with pT > 20 GeV and |η| < 4.5. Suchan event topology is optimal for estimation of the soft-termsystematic uncertainties because only the muons and the softterm contribute to the Emiss

T . In principle the methods arevalid in event topologies with any jet multiplicity, but theZ → µµ+ ≥1-jet events are more susceptible to jet-relatedsystematic uncertainties.

8.1.1 Evaluation of balance between the soft term and thehard term

The primary or “balance” method exploits the momentumbalance in the transverse plane between the soft and hardterms in Z → ℓℓ events, and the level of disagreementbetween data and simulation is assigned as a systematicuncertainty.

The E miss,softT is decomposed along the p hard

T direction.The direction orthogonal to p hard

T is referred to as the per-

123

NEW!

https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2017-024


Mono-H(𝛾𝛾)

12 [GeV]Z'm

600 800 1000 1200 1400

[GeV

]0 A

m

200

250

300

350

400

450 thσ/

obs

σ95

% C

L ob

serv

ed li

mit

on

1−10

1


), Z'-2HDM, Dirac DMγγmono-H( = 100 GeVχm = 0.8,

Z'g = 1.0, βtan

= 1thσ/obsσExpected

σExpected + 1 σExpected - 1

hm

+

0A

m =

Z'

m

thσ/

obs

σ95

% C

L ob

serv

ed li

mit

on

1−10

1

[GeV]BZ'm

10 210 310

[GeV

]χ

m

1

10

210

310 thσ/

obs

σ95

% C

L ob

serv

ed li

mit

on

1−10

1


, Dirac DMB

), Z'γγmono-H( = 1χg = 1/3, qg = 0.3, θsin

= 1thσ/obsσExpected

σExpected + 1 σExpected - 1

χm = 2

Z'm

• Non-resonant BG (w/o H): 𝛾𝛾, 𝛾+jets, V𝛾, V𝛾𝛾.

• Non-resonant BG (w/ H): Z(vv)H(𝛾𝛾).

• Likelihood fit in 105 < m𝛾𝛾 < 160 GeV using analytic functions.

• Double-sided Crystal Ball is used for the signals & Z(vv)H(𝛾𝛾).

• Consistent with the SM predictions. [GeV]γγm 110 120 130 140 150 160

Dat

a-Bk

g

5−05

[GeV]γγm 110 120 130 140 150 160

Eve

nts

/ GeV

024

6

8

1012

=125.09 GeV)h

m=1, µh (

= 200 GeVBZ'm = 1 GeV, χm

= 1/3, Dirac DMq

g = 0.3, θ, sinBZ'

Non-resonant bkg

Non-resonant bkg + h

Total

ATLAS Preliminary-1 = 13 TeV, 36.1 fbs

NEW!


Mono-H(𝛾𝛾)

13

[GeV]Hm260 270 280 290 300 310 320 330 340 350

[fb]

γγ →

h B

R×

χχ

h

→pp

σ

95%

CL

limit

on

1

10

210

310

410 ObservedExpected

σ 1±Expected σ 2±Expected

BR× thσ


, Heavy scalar, Scalar DMχχ) + γγ h(→ H →pp ) = 100%χχ h→ = 60 GeV, BR (H χm

[GeV]γγm 110 120 130 140 150 160

Eve

nts

/ GeV

0

10

20

30

40

50=125.09 GeV)

hm=1, µh (

= 275 GeVHm = 60 GeV, χmHeavy scalar, Scalar DM

Non-resonant bkg


Total


missT

High-E

[GeV]γγm 110 120 130 140 150 160

Dat

a-Bk

g

20−

020

[GeV]γγm 110 120 130 140 150 160

Eve

nts

/ GeV

020406080

100120140160180

=125.09 GeV)h

m=1, µh (


Non-resonant bkg


Total


missTIntermediate-E

[GeV]γγm 110 120 130 140 150 160

Dat

a-Bk

g

50−

050

[GeV]γγm 110 120 130 140 150 160

Eve

nts

/ GeV

0100200300400500600700 =125.09 GeV)

hm=1, µh (


Non-resonant bkg


Total


Different-Vertex

[GeV]γγm 110 120 130 140 150 160

Dat

a-Bk

g

50−

050

[GeV]γγm 110 120 130 140 150 160

Eve

nts

/ GeV

02468

10121416

310×

=125.09 GeV)h

m=1, µh (


Non-resonant bkg


Total


Rest

[GeV]γγm 110 120 130 140 150 160

Dat

a-Bk

g

200−0

200

• The same approach for the other 4 categories as well.

• Consistent with the SM prediction.

NEW!


Mono-H(bb)

14

Eve

nts

/ b

in

1

10

210

310

410

510

610

710

810

910DataZ+jets

+ single topttW+jetsDibosonSM VhMultijetBackground UncertaintyPre-fit Backgroundmono-h Z’-2HDM

= 600 GeVA

= 1400 GeV, mZ’m = 4.74 fbSignalσ

ATLAS Preliminary-1 = 13 TeV , 36.1 fbs

SR : 0 lepton

1 b-tag

0

[GeV]missTE

200 300 400 500 600 700 800

Data

/SM

0.95

1

1.050

Eve

nts

/ b

in

1

10

210

310

410

510

610

710DataZ+jets


= 600 GeVA



SR : 0 lepton

2 b-tags

0

[GeV]missTE

200 300 400 500 600 700 800

Data

/SM

0.8

1

1.20

• Advantage due to high BR(H→bb) ~ 58%.

• 8 signal regions categorized with the b-jet multiplicity (1 or 2 b-tags) & ETmiss bins.

• Small-R jets (R=0.4) are considered for ETmiss < 500 GeV & large-R jets (R=1.0) for the last bin.

ATLAS-CONF-2017-028

• Veto on 𝜏-leptons & additional b-jets. Some angular cuts for the resolved (using small-R jets) categories & some HT requirements to reduce BGs.

• Main BGs are Z+jets, W+jets, Top, diboson & Z(vv)H(bb).

• Normalization of Z+jets taken from 2-lepton CR, whereas 1-lepton CR is used for W+jets & Top.

NEW!



Mono-H(bb)

15

Eve

nts

/ 5

Ge

V

0

100

200

300

400

500

600

700

800

900 DataSM VhDiboson

+ single topttZ+jetsW+jetsMultijetBackground UncertaintyPre-fit Backgroundmono-h Z’-2HDM

= 600 GeVA



SR (Resolved) : 0 lepton

< 200 GeVmiss

T150 GeV < E

2 b-tags

0

[GeV]jjm50 100 150 200 250

Da

ta/S

M

0.5

1

1.50

Eve

nts

/ 5

GeV

0

50

100

150

200

250

300

350

400

450DataSM VhDiboson

+ single topttZ+jetsW+jetsMultijetBackground UncertaintyPre-fit Backgroundmono-h Z’-2HDM

= 600 GeVA




< 350 GeVmiss

T200 GeV < E

2 b-tags

0

[GeV]jjm50 100 150 200 250

Da

ta/S

M

0.5

1

1.50

Eve

nts

/ 1

0 G

eV

0

10

20

30

40

50

60

70

80

90

100Data

SM VhDiboson

+ single topttZ+jets

W+jetsBackground UncertaintyPre-fit Backgroundmono-h Z’-2HDM

= 600 GeVA




< 500 GeVmiss

T350 GeV < E

2 b-tags

0

[GeV]jjm50 100 150 200 250

Data

/SM

0.5

1

1.50

Eve

nts

/ 2

0 G

eV

0

5

10

15

20

25

30Data

SM VhDiboson

+ single topttZ+jets

W+jetsBackground UncertaintyPre-fit Backgroundmono-h Z’-2HDM

= 600 GeVA



SR (Merged) : 0 lepton

> 500 GeVmissTE

2 b-tags

0

[GeV]Jm50 100 150 200 250

Da

ta/S

M

0.5

1

1.50

• mjj or mJ is used as the final discriminant.


• Main systematics: b-tagging, V+jets modeling, SM VH normalization.

NEW!


Mono-H(bb)

16

[GeV]Z'm10 210 310 410

[GeV

]χ

m

1

10

210

310

410ATLAS

-1 = 13 TeV, 3.2 fbsVector Mediator

=1/3q=1, gχ

gZ'=m

Z')=0.3, gθsin(

Observed 95% CL LimitsExpected 95% CL Limits

σ1±Expected σ2±Expected

[GeV]Z'm500 1000 1500 2000 2500

[GeV

]A

m

300

400

500

600

700

800

900

1000h

- m

Z' =

mA

Kinematic lim

it: m

Observed limitσ1±Expected limit

-1 = 13 TeV, 3.2 fbs-1 = 8 TeV, 20.3 fbs


Mono-h(bb), All limits at 95 % CLZ'-2HDM simplified model

= 100 GeVχ

= 0.8, mZ

= 1, gβtan

• Mono-H(bb) has the best sensitivity among the Mono-H channels.

• For Z′-2HDM, the exclusion curves are shown with the ATLAS/CMS benchmark parameters of tan 𝛽=1, gZ=0.8, m𝜒=100 GeV.

• For vector mediator models, mZ′ < ~700 GeV is excluded for g𝜒=1, gq=1/3, sin𝜃=0.3, gZ′=mZ′.

Phys. Lett. B 765 (2016) 11

2015 Data

(3.2 fb-1 )

NEW!

http://www.sciencedirect.com/science/article/pii/S0370269316306980




Summary

17

• Presented the latest results of mono-W/Z and mono-H searches at the ATLAS experiment.

• These searches are direct probes of BSM interactions with the vector bosons/Higgs boson and the dark matter.

• No deviation from the Standard Model predictions at the moment.

• New models (e.g. an extra scalar with H mixing, etc.) are proposed by the LHC DM WG, and are being considered for the near future.

• More results using the full 2015+2016 dataset will follow.

backups


D2 for Boosted-V Tagging

19

154 Page 4 of 47 Eur. Phys. J. C (2016) 76 :154

the distance between each constituent and any of these axesis RaN ,i . In the above functions, the sum is performed overthe constituents i in the jet J , such that the normalisationfactor τ0 (Eq. 5a) is equivalent to the magnitude of the jet pTmultiplied by the β-exponentiated jet radius.

Recent studies [29] have shown that an effective alterna-tive axis definition can increase the discrimination power ofthese variables. The ‘winner-takes-all’ axis uses the directionof the hardest constituent in the exclusive kt subjet insteadof the subjet axis, such that the distance measure #Ra1,ichanges in the calculation. The ratio of the N-subjettinessfunctions found with the standard subjet axes, τ21, and withthe ‘winner-takes-all’ axes, τwta

21 , can be used to generate thedimensionless variables that have been shown in particle-level MC to be particularly useful in identifying two-bodystructures within jets:

τ21 = τ2

τ1, τwta

21 = τwta2

τwta1

. (6)

Energy correlation ratios:The 1-point, 2-point and 3-point energy correlation func-

tions for a jet J are given by:

ECF0(β) = 1, (7a)

ECF1(β) =!

i∈J

pTi , (7b)

ECF2(β) =!

i< j∈J

pTi pT j (#Ri j )β , (7c)

ECF3(β) =!

i< j<k∈J

pTi pT j pTk (#Ri j#Rik#R jk)β , (7d)

where the parameter β is used to give weight to the angularseparation of the jet constituents. In the above functions, thesum is over the constituents i in the jet J , such that the 1-point correlation function Eq. (7b) is approximately the jetpT. Likewise, if one takes β = 2, it is noted that the 2-pointcorrelation functions are equivalent to the mass of a particleundergoing a two-body decay in collider coordinates.

An abbreviated form of these definitions can be writtenas:

e(β)2 = ECF2(β)

ECF1(β)2 , (8a)

e(β)3 = ECF3(β)

ECF1(β)3 . (8b)

These ratios of the energy correlation functions can beused to generate the dimensionless variable C (β)

2 [16], andits more recently modified version D(β)

2 [17,18], that havebeen shown in particle-level MC to be particularly useful inidentifying two-body structures within jets:

C (β)2 = e(β)3

(e(β)2 )2, (9a)

D(β)2 = e(β)3

(e(β)2 )3. (9b)

Values of β = 1 and 2 are studied here.

3 The ATLAS detector

The ATLAS detector [30] at the LHC covers nearly theentire solid angle around the collision point. It consists of aninner tracking detector surrounded by a thin superconductingsolenoid, electromagnetic and hadronic calorimeters, and amuon spectrometer incorporating three large superconduct-ing toroid magnets.

The inner-detector system (ID) is immersed in a 2 T axialmagnetic field and provides charged particle tracking in therange |η| < 2.5. A high-granularity silicon pixel detectorcovers the vertex region and typically provides three mea-surements per track. It is followed by a silicon microstriptracker, which usually provides four two-dimensional mea-surement points per track. These silicon detectors are com-plemented by a transition radiation tracker, which enablesradially extended track reconstruction up to |η| = 2.0. Thetransition radiation tracker also provides electron identifica-tion information based on the fraction of hits (typically 30 intotal) above a higher energy-deposit threshold correspondingto transition radiation.

The calorimeter system covers the pseudorapidity range|η| < 4.9. Within the region |η| < 3.2, electromag-netic calorimetry is provided by barrel and endcap high-granularity lead/liquid-argon (LAr) electromagnetic calorime-ters, with an additional thin LAr presampler covering |η| <1.8, to correct for energy loss in material upstream of thecalorimeters. For the jets measured here, the transverse gran-ularity ranges from 0.003 × 0.1 to 0.1 × 0.1 in #η × #φ,depending on depth segment and pseudorapidity. Hadroniccalorimetry is provided by a steel/scintillator-tile calorime-ter, segmented into three barrel structures within |η| < 1.7,and two copper/LAr hadronic endcap calorimeters. This sys-tem enables measurements of the shower energy depositionin three depth segments at a transverse granularity of typi-cally 0.1 × 0.1. The solid angle coverage is extended withforward copper/LAr and tungsten/LAr calorimeter modulesoptimised for electromagnetic and hadronic measurementsrespectively.

A muon spectrometer (MS) comprises separate trigger andhigh-precision tracking chambers measuring the deflection ofmuons in a magnetic field generated by superconducting air-core toroids. The precision chamber system covers the region|η| < 2.7 with three layers of monitored drift tubes, com-

123





τ21 = τ2

τ1, τwta

21 = τwta2

τwta1

. (6)



ECF0(β) = 1, (7a)

ECF1(β) =!

i∈J

pTi , (7b)

ECF2(β) =!

i< j∈J


ECF3(β) =!

i< j<k∈J




e(β)2 = ECF2(β)

ECF1(β)2 , (8a)

e(β)3 = ECF3(β)

ECF1(β)3 . (8b)




C (β)2 = e(β)3

(e(β)2 )2, (9a)

D(β)2 = e(β)3

(e(β)2 )3. (9b)







123





τ21 = τ2

τ1, τwta

21 = τwta2

τwta1

. (6)



ECF0(β) = 1, (7a)

ECF1(β) =!

i∈J

pTi , (7b)

ECF2(β) =!

i< j∈J


ECF3(β) =!

i< j<k∈J




e(β)2 = ECF2(β)

ECF1(β)2 , (8a)

e(β)3 = ECF3(β)

ECF1(β)3 . (8b)




C (β)2 = e(β)3

(e(β)2 )2, (9a)

D(β)2 = e(β)3

(e(β)2 )3. (9b)







123

𝛽=1 is considered


Boosted-Boson Tagging

20

[GeV]T

Jet p500 1000 1500 2000

@ 5

0% s

igna

l eff.

2D

1

1.5

2

2.5

3

Fit: fourth order polynomial

ATLAS Simulation Preliminary R=1.0 jetstanti-k

= 0.2)sub = 5%, Rcut

Trimmed (f| < 2.0, Z-jetsη|

(a) Z-jets, ‘Medium’.

[GeV]T

Jet p500 1000 1500 2000

@ 2

5% s

igna

l eff.

2D

0.8

1

1.2

1.4

1.6

Fit: second order polynomial


= 0.2)sub = 5%, Rcut


(b) Z-jets, ‘Tight’.

[GeV]T

Jet p500 1000 1500 2000

@ 5

0% s

igna

l eff.

2D

1

1.5

2

2.5

3



= 0.2)sub = 5%, Rcut

Trimmed (f| < 2.0, W-jetsη|

(c) W -jets, ‘Medium’.

[GeV]T

Jet p500 1000 1500 2000

@ 2

5% s

igna

l eff.

2D

0.8

1

1.2

1.4

1.6



= 0.2)sub = 5%, Rcut


(d) W -jets, ‘Tight’.

Figure 8: The D2 maximum cut value that yields a ‘Medium’ 50% signal eciency (left) and a ‘Tight’ 25% signaleciency (right), when combined with the calibrated mass window for Z-jets and W -jets as a function of calibratedjet pT.

15

[GeV]T

Jet p500 1000 1500 2000

@ 5

0% s

igna

l eff.

2D

1

1.5

2

2.5

3



= 0.2)sub = 5%, Rcut


(a) Z-jets, ‘Medium’.

[GeV]T

Jet p500 1000 1500 2000

@ 2

5% s

igna

l eff.

2D

0.8

1

1.2

1.4

1.6



= 0.2)sub = 5%, Rcut


(b) Z-jets, ‘Tight’.

[GeV]T

Jet p500 1000 1500 2000

@ 5

0% s

igna

l eff.

2D

1

1.5

2

2.5

3



= 0.2)sub = 5%, Rcut


(c) W -jets, ‘Medium’.

[GeV]T

Jet p500 1000 1500 2000

@ 2

5% s

igna

l eff.

2D

0.8

1

1.2

1.4

1.6



= 0.2)sub = 5%, Rcut


(d) W -jets, ‘Tight’.

Figure 8: The D2 maximum cut value that yields a ‘Medium’ 50% signal eciency (left) and a ‘Tight’ 25% signaleciency (right), when combined with the calibrated mass window for Z-jets and W -jets as a function of calibratedjet pT.

15


Mono-Z(ℓℓ)

21

2015 + 2016

(13.3 fb-1 )

Event SelectionExactly one ee or µµ pair

pT

(e/µ) > 30(20) GeV for leading (sub-leading) lepton

Selection High Mass Low Mass|mll mZ | < 15 GeV

Emiss

T

> 120 GeV > 90 GeV

R`` < 1.8

| (~p``T

, ~Emiss

T

) | > 2.7

|pmiss,jet

T

p``T

|/p``T

< 0.2

| (~Emiss

T

, jets) |> 0.4 > 0.7

pT

(jet) > 100 GeV pT

(jet) > 25 GeV

p``T

/mT

< 0.7 < 0.9

Number of b-jets = 0

Low Mass Signal Region ee µµData 220 236

Signals

ZH (mH = 125 GeV) with BF(H ! invisible)=100%

40.5 ± 1.2 ± 4.1 41.7 ± 1.2 ± 4.4Mono-Z (m = 1 GeV, m

med

= 10 GeV)

175 ± 24 ± 14 169 ± 21 ± 22

Mono-Z (m = 50 GeV, mmed

= 300 GeV)

43.7 ± 2.3 ± 2.8 49.1 ± 2.6 ± 4.2Backgrounds

qqZZ (MC-based)

95.0 ± 1.5 ± 5.8 102.1 ± 1.6 ± 8.0ggZZ (MC-based)

5.6 ± 0.1 ± 3.3 5.7 ± 0.1 ± 3.4WZ (Data-driven)

44.0 ± 1.1 ± 3.3 50.5 ± 1.2 ± 3.3Z(! ee, µµ)+jets (Data-driven)

23 ± 5 ± 11 16.9 ± 5.2 ± 6.7non-resonant-`` (Data-driven)

16.9 ± 2.8 ± 1.0 20.7 ± 3.4 ± 1.2fake-lepton (Data-driven)

0.18 ± 0.04 ± 0.03 0.36 ± 0.46 ± 0.08

t¯tV/VVV (MC-based)

0.44 ± 0.02 ± 0.06 0.43 ± 0.02 ± 0.06

Total background

185 ± 6 ± 13 196 ± 7 ± 12


Mono-H(𝛾𝛾)

22

[GeV]χmDM mass 1 10 210 310

]2D

M-n

ucle

on c

ross

sec

tion

[cm

47−10

45−10

43−10

41−10

39−10

37−10

35−10

33−10

31−10

29−10

90% CLSpin-independent

PandaX-IILUX2016

CRESST-II 2016superCDMS

= 1χ

g = 1/3, qg = 0.3, θsin, Dirac DMB, Z'χ χ) + γγ h(→pp PreliminaryATLAS

-1 = 13 TeV, 36.1 fbs


Mono-H(𝛾𝛾)

23

[GeV]BZ'm

0 500 1000 1500 2000

[fb]

γγ →

h B

R× χχ

h→

pp

σ95

% C

L lim

it on

1−10

1

10

210



BR× thσ


, Dirac DMB

, Z'χ χ) + γγ h(→pp = 1 GeVχ = 1, m

χg = 1/3, qg = 0.3, θsin

10 2101

10

[GeV]0Am200 300 400 500 600 700 800

[fb]

γγ →

h B

R× χχ

h→

pp

σ95

% C

L lim

it on

3−10

2−10

1−10

110

210

310

410

510



BR× thσ


, Z'-2HDM, Dirac DMχ χ) + γγ h(→pp = 1 TeVZ'm = 100 GeV, χm = 0.8,

Z'g = 1.0, βtan

200 220 240 260 280 300 320 340

1


Mono-H(bb)

24

Appendix

q

qZ

0 h

A

Figure 4: Diagram for production of DM particles in association with a Higgs boson in a simplified model wherea Z0 decays to a Higgs boson h and a pseudoscalar A.

Table 3: A summary of the main analysis selection criteria. The notation pT(A, B) used here is defined as the vectorsum of the pT for the objects A and B. For detailed descriptions of the selection criteria, please refer to the textbody.

Region SR 1µ-CR 2`-CRTrigger Emiss

T EmissT Single lepton

Exactly two e or µLeptons No e or µ Exactly one µ 83 GeV < mee < 99 GeV

71 GeV < mµ±µ < 106 GeV

Resolved

EmissT 2 [150, 500] GeV pT (µ, Emiss

T ) 2 [150, 500] GeV pT (`, `) 2 [150, 500] GeVpmiss

T > 30 GeV (1 b-tag only) pT (µ, pmissT ) > 30 GeV –

minh~Emiss

T , ~p jetT

i> /9 min

h~Emiss

T , ~p jetT

i> /9 –

~Emiss

T , ~p missT

< /2

~Emiss

T , ~p missT

< /2 –

– – EmissT

Pjets,leptons pT

1/2< 3.5 GeV1/2

Number of central small-R jets 2Leading Higgs candidate small-R jet pT > 45 GeV

HT,2 jets > 120 GeV for 2 jets, HT,3 jets > 150 GeV for > 2 jets~Emiss

T , ~pT,h

> 2/3

Veto on -leptonsR~p jet 1

h , ~pjet 2h

< 1.8

Veto on events with > 2 b-tagsSum of pT of two Higgs candidate jets and leading extra jet > 0.63 HT,all jets

b-tagging : one or two small-R calorimeter jetsFinal discriminant = Dijet mass

Merged

EmissT > 500 GeV pT (µ, Emiss

T ) > 500 GeV pT (`, `) > 500 GeVpmiss

T > 30 GeV pT (µ, pmissT ) > 30 GeV –

minh~Emiss

T , ~p jetT

i> /9 min

h~Emiss

T , ~p jetT

i> /9 –

~Emiss

T , ~p missT

< /2

~Emiss

T , ~p missT

< /2 –

Number of large-R jets 1Veto on -lepton not associated to large-R jet

Veto on b-jets not associated to large-R jetHT -ratio selection (<0.57)

b-tagging : one or two ID track jets matched to large-R jetFinal discriminant = Large-R jet mass

12


Mono-H(bb)

25

Eve

nts

/ b

in

1

10

210

310

410

510

610

710DataZ+jets


= 600 GeVA



SR : 0 lepton

2 b-tags

0

[GeV]missTE

200 300 400 500 600 700 800

Da

ta/S

M0.8

1

1.20

Figure 2: Distribution of EmissT in the SR with two b-tags. The upper panel shows a comparison of data to the SM

expectation before (dashed lines) and after the fit with no signal included (solid histograms). The bottom paneldisplays the ratio of data to SM expectations after the fit, with its systematic uncertainty considering correlationsbetween individual contributions indicated by the shaded band. The expected signal from a representative Z0-2HDMmodel with (mZ0 ,mA) = (1.4 TeV, 0.6 TeV), assuming a production cross-section of 4.74 fb, is also shown (long-dashed line). The rightmost bin includes overflows.

[GeV]Z'm500 1000 1500 2000 2500

[GeV

]A

m

300

400

500

600

700

800

900

1000

h -

mZ'

= m

A

Kinematic lim

it: m

Observed limitσ1±Expected limit

-1 = 13 TeV, 3.2 fbs-1 = 8 TeV, 20.3 fbs


Mono-h(bb), All limits at 95 % CLZ'-2HDM simplified model

= 100 GeVχ

= 0.8, mZ

= 1, gβtan

Figure 3: CLs exclusion contours at 95% confidence level for the Z0-2HDM scenario in the (mZ0 ,mA) plane fortan = 1, gZ0 = 0.8, and m = 100 GeV. The observed limits (solid line) are consistent with the expectation underthe SM-only hypothesis (dashed line) within uncertainties (filled band). Also shown are the observed limits fromprevious ATLAS results at

ps = 8 TeV (long-dashed line) [27] and at

ps = 13 TeV (dash-dotted line) [29].

10

Eve

nts

/ b

in

1

10

210

310

410

510

610

710

810

910DataZ+jets


= 600 GeVA



SR : 0 lepton

1 b-tag

0

[GeV]missTE

200 300 400 500 600 700 800

Da

ta/S

M

0.95

1

1.050

Figure 5: Distribution of EmissT in the SR with one b-tag. The upper panel shows a comparison of data to the SM

expectation before (dashed lines) and after the fit with no signal included (solid histograms). The bottom paneldisplays the ratio of data to SM expectations after the fit, with its systematic uncertainty considering correlationsbetween individual contributions indicated by the shaded band. The expected signal from a representative Z0-2HDMmodel with (mZ0 ,mA) = (1.4 TeV, 0.6 TeV), assuming a production cross-section of 4.74 fb, is also shown (long-dashed line). The rightmost bin includes overflows.

[GeV]200 300 400 500 600 700 800

Frac

tion

of T

otal

Bac

kgro

und

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

-µ tt-µW+jets

+µ tt+µW+jets

ATLAS Simulation Preliminary = 13 TeVs

1 Lepton, 1 b-tags CR vs. W+jetstt Charge Separationµ

missTE

200 300 400 500 600 700 800

Frac

tion

of T

otal

Bac

kgro

und

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

-µ tt-µW+jets

+µ tt+µW+jets

ATLAS Simulation Preliminary = 13 TeVs

1 Lepton, 2 b-tags CR vs. W+jetstt Charge Separationµ

[GeV]missTE

Figure 6: Fraction of the tt and W+jets processes to the total sum of backgrounds as a function of EmissT for events

with one and two b-tags in the 1µ-CR. While the former is symmetric in the muon charge, the latter shows anasymmetry which allows to separate the two processes. The rightmost bin includes overflows.

13


Mono-H(bb)

26

Table 2: Upper limits at 95% confidence level on the visible cross-section vis,h+DM of h+DM events. The observedobs

vis,h+DM is consistent with the expectation expvis,h+DM within uncertainties. Also shown are the A " values to

reconstruct and select an event in the same EmissT bin as generated.

Range in obsvis,h+DM exp

vis,h+DM A "Emiss

T /GeV [fb] [fb] %[150, 200) 19.1 18.3+7.2

5.1 15[200, 350) 13.1 10.5+4.1

2.9 35[350, 500) 2.4 1.7+0.7

0.5 40[500,1) 1.7 1.8+0.7

0.5 55

In summary, a search for DM produced in association with a Higgs boson in final states with EmissT and a

bb pair from the h! bb decay was conducted using 36.1 fb1 of pp collisions atp

s = 13 TeV collectedby the ATLAS detector at the LHC. The results are in agreement with SM predictions, and a substantialfraction of parameter space of a representative Z0-2HDM model is excluded, significantly improving uponprevious results. Stringent limits are placed on the visible cross-section of non-SM events with a Higgsboson decaying as h! bb and large Emiss

T without extra model assumptions.

We thank CERN for the very successful operation of the LHC, as well as the support sta↵ from ourinstitutions without whom ATLAS could not be operated eciently.

We acknowledge the support of ANPCyT, Argentina; YerPhI, Armenia; ARC, Australia; BMWFW andFWF, Austria; ANAS, Azerbaijan; SSTC, Belarus; CNPq and FAPESP, Brazil; NSERC, NRC and CFI,Canada; CERN; CONICYT, Chile; CAS, MOST and NSFC, China; COLCIENCIAS, Colombia; MSMTCR, MPO CR and VSC CR, Czech Republic; DNRF and DNSRC, Denmark; IN2P3-CNRS, CEA-DSM/IRFU, France; GNSF, Georgia; BMBF, HGF, and MPG, Germany; GSRT, Greece; RGC, HongKong SAR, China; ISF, I-CORE and Benoziyo Center, Israel; INFN, Italy; MEXT and JSPS, Japan;CNRST, Morocco; FOM and NWO, Netherlands; RCN, Norway; MNiSW and NCN, Poland; FCT, Por-tugal; MNE/IFA, Romania; MES of Russia and NRC KI, Russian Federation; JINR; MESTD, Serbia;MSSR, Slovakia; ARRS and MIZŠ, Slovenia; DST/NRF, South Africa; MINECO, Spain; SRC andWallenberg Foundation, Sweden; SERI, SNSF and Cantons of Bern and Geneva, Switzerland; MOST,Taiwan; TAEK, Turkey; STFC, United Kingdom; DOE and NSF, United States of America. In addition,individual groups and members have received support from BCKDF, the Canada Council, CANARIE,CRC, Compute Canada, FQRNT, and the Ontario Innovation Trust, Canada; EPLANET, ERC, ERDF,FP7, Horizon 2020 and Marie Skłodowska-Curie Actions, European Union; Investissements d’AvenirLabex and Idex, ANR, Région Auvergne and Fondation Partager le Savoir, France; DFG and AvH Found-ation, Germany; Herakleitos, Thales and Aristeia programmes co-financed by EU-ESF and the GreekNSRF; BSF, GIF and Minerva, Israel; BRF, Norway; CERCA Programme Generalitat de Catalunya,Generalitat Valenciana, Spain; the Royal Society and Leverhulme Trust, United Kingdom.

The crucial computing support from all WLCG partners is acknowledged gratefully, in particular fromCERN, the ATLAS Tier-1 facilities at TRIUMF (Canada), NDGF (Denmark, Norway, Sweden), CC-IN2P3 (France), KIT/GridKA (Germany), INFN-CNAF (Italy), NL-T1 (Netherlands), PIC (Spain), ASGC(Taiwan), RAL (UK) and BNL (USA), the Tier-2 facilities worldwide and large non-WLCG resource pro-viders. Major contributors of computing resources are listed in Ref. [78].

11


Mono-H(4ℓ)

27

[GeV]missTE

0 50 100 150 200 250 300

Even

ts/1

0 G

eV

7−106−105−104−103−102−101−101

10210310410510610710 Data

Vector: 200 GeVScalar: 300 GeV

4l→HZZ*Z+jetstt

νlνZHlννZHll

PreliminaryATLAS = 1 GeVχm

-113 TeV, 3.2 fb < 140 GeV4l110 < m

[GeV]medm0 500 1000 1500 2000

4l)

[fb]

→ Z

Z*→

BR(

H×

σ95

% C

L lim

it on

1

10

210sCLObserved sCLExpected

σ 1 ±σ 2 ±

PreliminaryATLAS = 1 GeVχScalar: m

-113 TeV, 3.21 fb < 140 GeV4l110 < m

[GeV]medm0 500 1000 1500 2000

4l)

[fb]

→ Z

Z*→

BR(

H×

σ95

% C

L lim

it on

1

10

210sCLObserved sCLExpected

σ 1 ±σ 2 ±

PreliminaryATLAS = 1 GeVχVector: m

-113 TeV, 3.21 fb < 140 GeV4l110 < m

ATLAS-CONF-2015-059

• Mono-H search using the 4ℓ decay. Limited by statistics, but offers a clean signature.


• Results are interpreted in terms of scalar/vector DM.

2015 Data

(3.2 fb-1 )


search for dark matter in events with a single boson and missing...

Documents