Sandra Horvat for the ATLAS Collaboration Higgs searches ... · Max-Planck-Institut f¨ur Physik...

Max-Planck-Institut fur Physik

(Werner-Heisenberg-Institut)

Sandra Horvat for the ATLAS Collaboration

Higgs searches with leptonsin ATLAS

LHC Days in Split • 2 - 7 October 2006 • Split, Croatia

Outline/ Lepton Reconstruction/ Searches: SM ( H → 4`, H → WW , H → ττ ), MSSM ( A/H → µµ ) / Outlook

Higgs searches in ATLAS - current status

The first evaluation of the Higgs discovery potential in ATLAS has

been published already in 1999 (Technical Design Report, TDR).

Since then:

The detector layout was modified, affecting detector performance.

ATLAS software is approaching the final detector description.

⇒ Necessary to update or improve the existing analyses.

Also, some new, promising channels have been spotted.

⇓A busy period has started (Computing System Commissioning, CSC):

Demonstrate the physics readiness by detailed studies of key channels.

Studies are mainly concentrated to the preparation for the first data

and the low-luminosity running period.

Final physics and performance notes to appear early next year.


Reconstructing Leptons in ATLAS


Why to look for leptonic signatures?

Leptons (≡ electrons and muons) are

most efficiently reconstructed objects in the detector,

with a very low fake rate

and with the highest momentum resolution.

Lepton reconstruction performance can be well understood from data,

using Z → e+e− and Z → µ+µ− processes as a reference.

Z → µµ: ∼ 40 000 events at 100 pb−1 (in few days - weeks)

- statistics similar to TEVATRON today.

⇓

Some of the cleanest Higgs signatures are those

having leptons as the final Higgs decay products.


Lepton reconstruction in ATLAS detector

EM

&H

AD

CA

LOR

IME

TE

RIN

NE

R D

ET

EC

TO

R (

B =

2 T

)

MU

ON

SP

EC

TR

OM

ET

ER

(<

B>

= 0

.4 T

)

Muons:

Muon spectrometer track

(MS: trigger + drift-tube chamb.),

can be combined with the

inner detector track.

(ID: silicon detector + TRT)

Low-pT muons: ID track

combined with MS hits.

Electrons:

Shower-shape analysis in the

fine-granularity calorimeter,

clusters are always matched

with the ID track.

Low-pT electrons:

ID combined with clusters.

Isolation criteria (given by the calorimeter or by the inner detector):

suppressing leptons which come from jets.


Lepton reconstruction efficiency

20 2040 40pT (GeV)

60pT (GeV)

60|eta||eta|

0

0.2

0.4

0.6

0.8

1.0

0 00.5 0.51 11.5 1.52 22.5 2.50

0.2

0.4

0.6

0.8

1.0electron efficiencymuon efficiency

0

0.2

0.4

0.6

0.8

1.0

0

0.2

0.4

0.6

0.8

1.0

high− and low−pThigh−pT high−pT

high− and low−pThigh− and low−pThigh−pT

high− and low−pThigh−pT

phi

eta=0

eta=2.5

eta=1.5 eta=1.0

etaeta

phi

<efficiency> = 96%Muon reconstruction efficiency Electron reconstr. efficiency

<efficiency> = 90%

−3 −2 −1 0 1 2 3

−1

−2

−3

0

1

2

3

−3 −2 −1 0 1 2 3

−1

−2

−3

0

1

2

31.00

0.98

0.96

0.94

0.92

0.90

0.88

0.86

0.840.82

0.80

1.00

0.98

0.96

0.94

0.92

0.90

0.88

0.86

0.840.82

0.80

Low-pT algorithms help to improve the reconstruction efficiency.

Electrons more difficult than muons (Brehmsstrahlung, dead material).


Lepton momentum resolution

0

10

20

30

40

1 10 10

muon spectrometer (MS)inner detector (ID)combined (MS+ID)

Muon momentum resolution:

2 103

−3pseudorapidity,

p_T

reso

lutio

n (%

)

η

p_T

reso

lutio

n (%

)

p_T (GeV)

p_T

reso

lutio

n (%

)

p_T (GeV)

−2 −1 0 1 2 3

10 20 30 40 50

electrons

muons (MS + ID)

60

muons (MS + ID)

electrons

70 80 90 100

2

2

4

4

6

6

8

8

10

10

eta=0

eta=2.5

eta=1.5 eta=1.0

(TDR: < ε >=95%, pT -res.=3%) → Performance now degraded w.r.t. TDR.

Reconstruction algorithms still to undergo slight improvements.


Leptonic Higgs Signatures


Menu of leptonic Higgs decays

Standard Model Higgs boson:

bb

γγ

WW

ZZ

tt

Purely leptonic decay:

H → ZZ (∗) → (`+`−)(`+`−)

Leptons + missing transverse energy:

H →W +W− → (`+ν)(`−ν),VBF: qqH,H →W +W− → (`+ν)(`−ν)VBF: qqH,H → τ+τ− → (`+νν)(`−νν)

Heavy neutral MSSM Higgs boson:

10−4

10−3

10−2

10−1

50 100 150 200 250 300 350 400 450 500

M A (GeV)

BR

H

ττ

µµ

bbtt

tanβ = 3

tanβ = 30

MHMAX Scenario

Purely leptonic decay:

(bb)H/A→ µ+µ−,

most promising after the ττ-mode.

At large tanβ:

decays into WW, ZZ and γγ are suppressed.


H → ZZ (∗) → (`+`−)(`+`−)

Highly sensitive to the lepton reconstruction efficiency, ε4` = (ε`)4.

ε4µ=0.85, ε4e=0.66

+

−

+

−

l

l

lZ

H Z q

q

Z

Z

+

−

+

−

gq

q

tt

irreduciblereducible

l

*

l

l

l

l

+gg

b

b

Z−l

l

l

l

W

Wb

b

l

l0 50 100 150 200

m(4l) / GeV

nr.e

vent

s at

30/

fb no cuts

250 3000

50

100

150

200

250

300

350


H → ZZ (∗) → (`+`−)(`+`−)


ε4µ=0.85, ε4e=0.66

+

−

+

−

l

l

lZ

H Z q

q

Z

Z

+

−

+

−

gq

q

tt


l

*

l

l

l

l

+gg

b

b

Z−l

l

l

l

W

Wb

b

l

l0 50 100 150 200

m(4l) / GeV

nr.e

vent

s at

30/

fb

after lepton isolation

250 3000

5

10

15

20

25

30

Lepton isolation (cut on EmaxT (∆R)) optimized separately for

electrons and muons - rejects most of the tt and Zbb events.

l

0.7 0.8 0.9 10.7 0.8 0.9 1

100

200

300

400

500

0

50

100

150

200

250

300

350

400

signal efficiency

back

grou

nd re

ject

ion

fact

or

signal efficiency

back

grou

nd re

ject

ion

fact

or

PSfrag replacements

∆R = 0.4

∆R = 0.24e− events 4µ− events

∆R = 0.45

ET (∆R = 0.2) < 6 GeVET (∆R = 0.4) < 9 GeV


H → ZZ (∗) → (`+`−)(`+`−)


ε4µ=0.85, ε4e=0.66

+

−

+

−

l

l

lZ

H Z q

q

Z

Z

+

−

+

−

gq

q

tt


l

*

l

l

l

l

+gg

b

b

Z−l

l

l

l

W

Wb

b

l

l0 50 100 150 200

nr.e

vent

s at

30/

fb

m(4l) / GeV

after all cuts

250 300

2

4

6

8

10

12

14

16

18

20

Lepton isolation (cut on EmaxT (∆R)) optimized separately for

electrons and muons - rejects most of the tt and Zbb events.

l

0.7 0.8 0.9 10.7 0.8 0.9 1

100

200

300

400

500

0

50

100

150

200

250

300

350

400

signal efficiency

back

grou

nd re

ject

ion

fact

or

signal efficiencyba

ckgr

ound

reje

ctio

n fa

ctor

PSfrag replacements

∆R = 0.4

∆R = 0.24e− events 4µ− events

∆R = 0.45

ET (∆R = 0.2) < 6 GeVET (∆R = 0.4) < 9 GeV

Additional cuts: primary vertex, Z -mass, Z ∗-mass, EmissT , p4`

T ;

optimized for the highest signal significance.


H → ZZ (∗) → (`+`−)(`+`−): Results for 30 fb−1

100 100 100150 150 150200 200 200250 250 250300 300 300350 350 350400 400 40000

22

44

66

88

1010

0

1

2

3

4

5

6

7

8

ZZZbbtt

sig + bckZZZbbtt

sig + bck

ZZZbbtt

sig + bck

m4ℓ (GeV/c2)

even

ts/

(2GeV

/c2

)

m4ℓ (GeV/c2)

even

ts/

(2GeV

/c2

)

m4ℓ (GeV/c2)

even

ts/

(2GeV

/c2

)

mH=130 GeV/c2 mH=180 GeV/c2 mH=280 GeV/c2

FULL SIMULATION mH =130 GeV mH =160 GeV mH =180 GeV mH =280 GeV

(δm = ±5 GeV) (δm = ±6 GeV) (δm = ±7 GeV) (δm = ±20 GeV)

Nsignal (gg+VBF) 21.5±0.1 26±1 28.1±0.3 67.4±0.1

Nqq→ZZ (×1.3 f. gg → ZZ) 11.3±0.3 11.4±0.3 27.3±0.5 40.4±0.6

NZbb 2±2 2±2 1±1 0±2

Ntt 0±0.4 0±0.4 0.5±0.4 0±0.4

Significance (no K) 5.0±0.3 5.5±0.5 4.7±0.2 8.8±0.4

L for 5σ discovery 30 fb−1 25 fb−1 37.5 fb−1 11 fb−1

Differences between muon and electron reconstruction ⇒FULL SIMULATION H → 4e H → 4µ H → 2e2µ total

Significance mH =130 GeV/c2 1.9 2.6 3.2 5.0


H → ZZ (∗) → (`+`−)(`+`−): Ensemble test

Acctual 4`-mass distribution at 30 fb−1 will look more like this:

98

100 150 200 250 300 350 4000

2

4

6

8

10

12

14

16

Entries Data subsample nr.5

(if we’re lucky)

signal

sig + bck

m4l (GeV/c2)

even

ts/

(2GeV

/c2

)

126

100 150 200 250 300 350 400

Entries

(bad luck)

Data subsample nr.17

0

2

4

6

8

10

signal

sig + bck

m4l (GeV/c2)

Small number of entries ⇒variable bin width instead of

equidistant bins for the fit of

the (S+)B-functions.

Ensemble test of the fit performance (60 subsamples, 25 fb−1 each):

fb (mk) = Nb · α2(mk − ε)e−α(mk−ε); fs (mk) = Ns√2πσ

· e−(mk−µ)2

2σ2

100 140 180 2200

2

4

6

8

10

12

backgroundhypothesis

signal+bckgr.hypothesis

variable bin widths,constant #entries/bin

m4ℓ (GeV/c2)

entr

ies

/(2

GeV

/c2

)

fit results remark

Ngood fits 54 max. 60

<Ns - Ntrues > 2 <Ntrue

s >=23

<Nb - Ntrueb > 3 <Ntrue

b >=86

<χ2

b−χ2s+b

χ2s+b> 1.6 hypothesis test

<Signf.> ”δm”-signf.

= Ns√Var(Ns )

2.9±0.6 4.1±0.3


H → W +W− → (`+ν)(`−ν)

Dominant backgrounds Rejection cuts

qq/gg → WW irreduciblett → WbWb M`` <300 GeV, b-jet veto

Z → `` M`` 6∈(82,98) GeV, EmissT >30 GeV

W + jets no hard jets

Azimuthal angle between the leptons, ∆Φ`` <1.5

(leptons emitted in the same direction, due to spin correlations).

Transverse mass, MT =p

2p``T · Emiss

T · (1− cos θ): MT ∈(50,MH+10) GeV.

Counting experiment: excess of signal over the expected background

(good knowledge of background distributions needed).

Recently: comparison of generators (PYTHIA, MC@NLO, SHERPA, ALPGEN)

- good agreement observed


VBF: qqH , H → W +W− → (`+ν)(`−ν)

New promising signature from the Vector Boson Fusion (VBF) process.

PSfrag

W, Z

H

ℓ−

ν

ℓ+

ν

W, Z

qq

q q

W+

W−

Typical topology:

two forward jets

leptons inbetween

rapidity gap

(no central jets)

50% improvement of the signal significance with Neural Network:

jj

Neural Net Output

Significance = 16.2

L=30/fb

η∆

0 0.2 0.4 0.6 0.8 1

10−3

0

10

1 2

−2

3 4 5 6 7 8 9 10

10−1

1

10

0

10

0.005

0.01

2

0.015

0.02

0.025

0.03

103

(GeV)llM0 100 200 300 400 500

Arbi

trary

Uni

ts

0

0.05

0.1

0.15

0.2

0.25

0.3 =160GeVH, MµSignal, e

tPythia tEW WWQCD WWEW Z+jetsQCD Zjj

llφ∆0 0.5 1 1.5 2 2.5 30

0.01

0.02

0.03

0.04

0.05

(GeV)TM

B.Mellado et al.

0 100 200 300 400 500

Arbi

trary

Uni

ts

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

Cros

s−se

ctio

n(fb

)


VBF: qqH , H → τ+τ− → (`+νν)(`−νν)

This channel is only visible in the VBF mode:

without the jet-cuts, the contribution from Z → ττ is too large.

cut number0 1 2 3 4 5 6 7 8 9 10 11

trig

ger Lepton−Cuts Jet−Cuts

all

12

Num

ber

of e

vent

s at

30/

fb

10

210

310

410

510

610

710

810

910g

q

q

b

t

t

b

W

W

tt

q g

q

q

Z

q

Zjj (QCD)

q

q

W

WWjj (QCD)

g

g

Z + jets, Z → ττ

WW + jets

Z + jets, Z → ee, µµtt

H(120GeV )→ ττ

nℓ>

1,

njet>

1

∆φ(2

ℓ)

Em

iss

T

xℓ

∆R

(2ℓ)

ηjet1×

ηjet2

<0

∆η(2

j)

mjj

∆φ(2

j)

centr

alje

tvet

o

mass

win

dow

number of central jets−10 −8 −6 −4 −2 0 2 4 6 8 10 0 200 400 600 800 1000 1200

0 11400

2 31600

4 51800

6 72000

8 9 10

even

ts

0

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

0

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

��

��

��

��

��

��

H(120GeV )→ ττ , ATLFAST

tt

Z + jets, Z → ee, µµ


WW + jets

H(120GeV )→ ττ , FULL SIM

∆η(2j)

even

ts

even

ts

mjj / GeV


VBF: qqH , H → τ+τ− → (`+νν)(`−νν)

Mass peak reconstruted by means of the collinear approximation:

mH � mτ , so products of τ-decays fly in the direction of τ-s.

Possible to calculate the four-momenta for τ-s.

lep

lep

missing ETν

ν

νν

ν

ν

ν

ν

τ

τ

0 20 40 60 80 100 120 140 160 180 2000

2

4

6

8

10

12

14

16

18

20

��

��

��

��

PSfrag replacements

Significance = 4.3

mττ / GeV


WW + jets

Z + jets, Z → ee, µµtt

H(120GeV )→ ττ

nr.ev

ents

(nor

mal

ized

to30/fb

)/

4GeV

0 40 80 120 160 200 0 2 4 6 81012141618 Subsample of 30/fb:

entr

ies

/ (4

GeV

)

mττ (GeV)

Similar sensitivity observed also in the semi-leptonic channel,

H → ττ → (`νν)(hadrons).


Remark on the VBF signatures

The searches in the VBF channels strongly rely on a good understanding

of the jet distributions:

Theory: underlying event uncertainties.Experiment:

Pile-up affects the rapidity gap between two forward jets.

Jet energy calibration (cross-section depends on the pjetT − cuts).

Comparison of different generators → hint for systematic uncertainties:

Delta Eta of two leading jets

I.Rottlaender

PT of central jets

I.Rottlaender PYTHIA 6.323, new showering model; BUG

PYTHIA 6.323, old showering model; BUG

PYTHIA 6.403, new showering model

PYTHIA 6.403, old showering model

HERWIG

Final evaluation of the generators and det. perf. to be done with data.(Z + jets: handle for the jet distributions;

reference for the Higgs VBF cross-section (same depend. on the pjetT -cuts).

V.A.Khoze et al., hep-ph/0207365v3


SM Higgs discovery potential

(GeV)HM100 120 140 160 180 200

Sign

ifica

nce

1

10

210

−1 L dt=30fb∫γγ→

TDR (1999) results

H 4l→ ZZ→H

bb→ttH,Hν l lν l l→ WW→H

Performance of an up-to-date detector is similar to the TDR.


SM Higgs discovery potential

(GeV)HM100 120 140 160 180 200

Sig

nific

ance

1

10

210

−1 L dt=30fb∫

qqWW→qqHττ qq→qqH

γγ→H 4l→ ZZ→H

bb→ttH,Hν lν l→ WW→H

Combined

K.Cranmer et al.

Performance of an up-to-date detector is similar to the TDR.

Vector boson fusion channels largely improve the discovery reach.


MSSM (bb)A/H → µ+µ−

The power of the lepton reconstruction performance:

The most promising A/H signature is the ττ-decay mode.

The µµ-decays are suppressed by the factorBR(A/H→µµ)BR(A/H→ττ) =

(mµ

mτ

)2= 1

282 .

However, the muons provide a much narrower Higgs resonance:

100invariant mass (GeV)

nr.e

ntrie

s no

rmal

ized

to 1

200 300 400 500 6000

0.05

0.10

0.15

0.20

0.25

0.30

µµA−>

collinear approximationττ −> (l )(hadrons)ννA−>


MSSM (bb)A/H → µ+µ−: Event selection

µ

µ

µ

µ

g

g b A/H

Associated production:Direct production:

+g

b

bg

q

g

b

bq µ

µA/H

A/H

Associated production with two b-jets

dominates at larger tanβ values (>10).

b-tag essential for the suppression

of the Z/γ∗ → µ+µ− background.

tt suppressed by the cuts on EmissT and pb−jet

T(b-jets more energetic than in signal).

(bb)A/H → µ+µ−

10-2

10-1

1

10

10 2

200 400 600mA (GeV)

σxB

R

(fb)

tan β = 3

tan β = 10

tan β = 30

totalassociated

Z/ *γ

g

µ

µZ

q

q

Q

Q

Z/ *γ1 400 000 fb

Drell−Yan

gg

t

t

W

W µ

µν

ν

b

b

6 950 fbt t

q

q

µ

µ

Z+jet940 000 fb


MSSM (bb)A/H → µ+µ−: Results for 30 fb−1

For the first time analysis performed using the full simulation:

Significance (tan β=30) 200 GeV 300 GeV 350 GeV 450 GeV

ATLFAST 6.5 1.9 1.5 0.7

FULL 5.4 1.7 – 0.4

m(2 mu) / GeV

bbA->mu mu tt->bb mu mu bbA+tt

100 200 300 400 500 600

5

0

10

15

20

1

2

3

4

5

6789

10

20

30

40

100 200 300 400 500 600 700 800 900

M.Schumacher

1000

−1 L dt=30fb∫

MHMAX scenario

bb h/H/A; h/H/A ττ

MA (GeV)

tanβ

MHMAX scenario

excluded by LEP (prel.)

bb h/H/A; h/H/A µµ

⇒ Only a factor of 2-3 worse than for the ττ decay mode.

Di-muon mode gives narrow resonances → separation of h/A/H peaks.


Outlook

The evaluation of the Higgs discovery potential for the key channels in

ATLAS approaches the final stage.

Leptonic signatures have a large contribution to the overall signal

significance, VBF channels are especially promising.

To Do List:

Fine-tune the reconstruction alghoritms.

Include the detailed information on the trigger performance.

Study the pile-up effects (especially for VBF).

Include the misalignment effects (mainly for the Higgs properties).

Prepare for the detector calibration and background estimation

with the first data.

...

The ATLAS detector is growing large, installation on schedule. →Pray for the first collisions.

Enjoy (the sleepless nights with) the real data.


Sandra Horvat for the ATLAS Collaboration Higgs searches ... · Max-Planck-Institut f¨ur Physik...

Documents

Transcript of Sandra Horvat for the ATLAS Collaboration Higgs searches ... · Max-Planck-Institut f¨ur Physik...