[PPT] - tb Heavy quarks and the Higgs: t b with ATLAS Tom Neep October 9, PowerPoint Presentation

SLIDE 1

Heavy quarks and the Higgs: t¯ tb¯ b with ATLAS

Tom Neep October 9, 2019

This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme under grant agreement no 714893 (ExclusiveHiggs)

1

SLIDE 2

The top quark

The top quark is the heaviest known

fundamental particle

It has a mass of ≈ 173 GeV. 40 times the

mass of the next heaviest quark!

Similar to that of a gold atom.
Discovered in 1995 at the Tevatron by the

CDF and D0 experiments

“Rediscovered” at the LHC in 2010
Unique amongst the quarks – decays

before hadronisation

Yukawa coupling of O(1). Some special

relationship with the Higgs?

2

SLIDE 3

Top quark pair production

Top quarks produced most often

in pairs

At the √s = 13 TeV around 90% of

t¯ t pairs are produced via gg → t¯ t and the remaining 10% by q¯ q → t¯ t

σt¯

t ≈ 830 pb (NNLO+NNLL QCD)

→≈ 10 t¯ t pairs /s at a luminosity

f 1034cm−2 s−1

q ¯ q ¯ t t ¯ t t

Large number of tops allows us to

make precise cross-section measurements

Many new physics models

enhance the t¯ t cross-section

Large number of t¯

t pairs allows us to measure t¯ t + X where X can be H, W, Z, γ, b¯ b and possibly one day even t¯ t

3

SLIDE 4

Top quark decay

The top quark decays nearly 100% of the time to a b-quark and a W-boson

t¯ t → W+W−b¯ b

t¯

t decays are therefore categorised based on how each of the two Ws decay

Three main channels
1. All-hadronic
2. Dilepton
3. Semi-leptonic
The t¯

t final state can include electrons, muons, taus, neutrinos (not detected) and jets (including b-jets).

We need to make use of the entire ATLAS

detector to make measurements!

all-hadronic electron+jets electron+jets muon+jets muon+jets tau+jets tau+jets

eµ eτ eτ

µτ µτ ττ

e+ cs ud

τ+ µ+

e– cs ud

τ– µ–

Top Pair Decay Channels

W d e c a y

eµ ee

µµ

d i l e p t

n

s

4

SLIDE 5

The ATLAS detector

5

SLIDE 6

b-tagging

JHEP 08 (2018) 89

b-tagging is crucial for top physics
Exploit large impact parameters,

seconday vertices and b → c decay chains

Information is combined using a

Boosted Decision Tree to identify jets containing b-hadrons

1 − 0.8 − 0.6 − 0.4 − 0.2 − 0.2 0.4 0.6 0.8 1

MV2c10 BDT output distribution

3 −

10

2 −

10

1 −

10 1

Event fraction

|<2.5 η >20 GeV, |

T

jet p

ATLAS Simulation t = 13 TeV, t s

b-jets c-jets Light-flavour jets

ǫb [%] light-jet mistag c-jet mistag 60 1550 35 70 380 12 77 135 6 85 35 3

6

SLIDE 7

NEW: t¯ t cross-section in eµ events

ATLAS-CONF-2019-041

The most accurate t¯

t cross-section measurements have been made in the eµ channel

This is a very clean channel with only small backgrounds
“Simple” technique, count the number of b-tagged jets

N1 = Lσt¯

tǫeµ2ǫb(1 − Cbǫb) + Nbkg 1

N2 = Lσt¯

tǫeµCbǫ2 b + Nbkg 2

L: Integrated luminosity σt¯

t: t¯

t cross-section ǫeµ: Efficiency for event to have one electron and one muon (≈ 1%) ǫb: Efficiency to tag and select a b-jet Cb: b-tagging correlation ≈ 1 Nbkg

1,2 : Number of background events

with 1/2 b-tags

Events 20 40 60 80 100 120 140 160 180 200 220

3

10 ×

ATLAS Preliminary

1

= 13 TeV, 36.1 fb s Data 2015+16 Powheg+PY8 t t Wt Z+jets Diboson Mis-ID lepton Powheg+PY8 Powheg+PY8 RadUp Powheg+PY8 RadDn aMC@NLO+PY8

b-tag

N 1 2 3 4 ≥ MC / Data 0.8 1 1.2

Stat. uncert.

7

SLIDE 8

NEW: t¯ t cross-section in eµ events

ATLAS-CONF-2019-041

Uncertainty source

∆σt¯

t/σt¯ t

(%) Data statistics 0.44 t¯ t mod. 0.97 Lept. 0.59 Jet/b 0.21 Bkg. 0.78 Analysis systematics 1.39 Integrated luminosity 1.90 Beam energy 0.23 Total uncertainty 2.40

Uncertainties are statistical,

systematic, luminosity and beam energy

The total uncertainty is dominated

by the luminosity uncertainty

t¯

t and background modelling are the next largest

8

SLIDE 9

NEW: t¯ t cross-section in eµ events

ATLAS-CONF-2019-041

Result: σt¯

t = 826.4 ± 3.6 ± 11.5 ± 15.7 ± 1.9 pb (2.4%)

Analysis has now been performed at 7, 8 and 13 TeV
All results are consistent with the SM (NNLO+NNLL QCD) prediction
The measurement is more precise than the prediction!

6 7 8 9 10 11 12 13 14 [pb]

t t

σ cross-section t Inclusive t

2

10

3

10

ATLAS Preliminary

+ b-tagged jets µ e

1

= 13 TeV, 36.1 fb s

1

= 8 TeV, 20.2 fb s

1

= 7 TeV, 4.6 fb s NNLO+NNLL (pp) Czakon, Fiedler, Mitov, PRL 110 (2013) 252004 uncertainties from PDF4LHC

S

α =172.5 GeV, PDF+

t

m [TeV] s 6 7 8 9 10 11 12 13 14

Ratio wrt PDF4LHC 0.9 0.95 1 1.05 1.1

NNPDF2.3 MSTW CT10 QCD scales only

[GeV]

pole t

m 164 166 168 170 172 174 176 178 180 182 ) [pb] t (t σ cross-section t Inclusive t 600 700 800 900 1000 1100 ATLAS Preliminary

1

= 13 TeV, 36.1 fb s CT14 NNLO+NNLL Czakon, Fiedler, Mitov, PRL 110 (2013) 252004

Some Birmingham involvement! EB chair: Miriam Watson Top cross-section convener: TN mpole

t

= 173.1 ± 1.0(exp.)+1.8

−2.1(theory)

GeV

9

SLIDE 10

t¯ t + X

ATLAS has also measured t¯

t production in association other particles

50 100 150 200 250 300 ) [GeV] γ (

T

p 0.75 0.875 1 1.125 Data / Pred. 200 400 600 800 1000 1200 1400 1600 Events ATLAS Preliminary

1

= 13 TeV, 139 fb s µ e Pre-Fit Data µ e γ t t γ t Other t γ Wt h-fake e-fake bkg. γ Prompt Uncertainty 82 84 86 88 90 92 94 96 98 100

[GeV]

ll

m

5 10 15 20 25 Events / 1.33 GeV ATLAS ATLAS

1

= 13 TeV, 36.1 fb s

1

= 13 TeV, 36.1 fb s 3L-Z-2b4j (pre-fit) 3L-Z-2b4j (pre-fit) Data Z t t WZ tZ tWZ Fake Leptons +X γ Uncertainty 0.4 0.6 0.8 1 1.2 1.4 1.6 W cross section [pb] t t 0.6 0.8 1 1.2 1.4 1.6 Z cross section [pb] t t Best fit 68% CL 95% CL NLO prediction

ATLAS

1

= 13 TeV, 36.1 fb s

t¯

t + γ and t¯ t + Z give very clean signals

Can start to measure differential

distributions

t¯

tW more challenging

Searches for t¯

tt¯ t ongoing, but will likely need more data for evidence

2 4 6 8 10

SM t t t t

σ /

t t t t

σ = µ 95% CL limit on Combined SS dilep. / trilep. Single lep. / OS dilep. ATLAS

1

= 13 TeV, 36.1 fb s (SM) t t t t σ 1 ± Expected σ 2 ± Expected Observed =1) µ Expected (

10

SLIDE 11

The Higgs and the top

It is important that we make the

most of the LHC and study the Higgs as comprehensively as possible

The top Yukawa coupling can be

probed through loops but also directly in Higgs production in association with top quarks (t¯ tH) t H ¯ t H t

all-hadronic electron+jets electron+jets muon+jets muon+jets tau+jets tau+jets

eµ eτ eτ µτ µτ ττ

e+ cs ud τ+ µ+ e– cs ud τ– µ–

Top Pair Decay Channels

W decay eµ ee µµ

d i l e p t

n

s

[GeV]

H

M

80 100 120 140 160 180 200

Higgs BR + Total Uncert

4

10

3

10

2

10

1

10 1

LHC HIGGS XS WG 2013

b b τ τ µ µ c c gg γ γ γ Z WW ZZ

The t¯

tH process can decay to a large number of different final

states. The more we measure the

better!

H → b¯

b is the dominant decay – can we measure t¯ tH(H → b¯ b)?

11

SLIDE 12

t¯ tH

Phys. Lett. B 784 (2018) 173
ATLAS observed t¯

tH production last year

Sensitivity comes mainly from the

H → γγ and multilepton channels (H → ττ and H → WW∗)

H → b¯

b not competitive, despite large branching ratio

The sensitivity of the t¯

tH(b¯ b) channel is limited by systematic uncertainties on the QCD t¯ tb¯ b background

SM ttH

σ /

ttH

σ 1 − 1 2 3 4

Total Stat. Syst. SM ATLAS

1

= 13 TeV, 36.1 - 79.8 fb s Total Stat. Syst. Combined )

0.19 0.21

± 0.18 , ± (

0.26 0.28

± 1.32 H (ZZ) t t < 1.77 at 68% CL ) γ γ H ( t t )

0.17 0.23

± ,

0.38 0.42

± (

0.42 0.48

± 1.39 H (multilepton) t t )

0.27 0.30

± ,

0.29 0.30

± (

0.40 0.42

± 1.56 ) b H (b t t 0.53 ) ± ,

0.28 0.29

± (

0.60 0.61

± 0.79

12

SLIDE 13

t¯ tH(bb)

Phys. Rev. D 97 (2018) 072016

Classification BDT output 1 − 0.8 − 0.6 − 0.4 − 0.2 − 0.2 0.4 0.6 0.8 1 Data / Pred.

0.5 0.75 1 1.25 1.5

Events / bin 50 100 150 200 250 ATLAS

1

= 13 TeV, 36.1 fb s Dilepton

1 4j ≥

SR Post-Fit Data H t t + light t t 1c ≥ + t t 1b ≥ + t t + V t t t Non-t Total unc. H (norm) t t

Measuring t¯

tH(bb) is extremely challenging

Final state with four b-jets – need to determine

which jets are from H → b¯ b and which are from t → Wb

Use MVA techniques to reconstruct the system

and to separate signal from background

Background is completely dominated by

t¯ tb(¯ b)

(reco BDT) [GeV]

Higgs bb

m 50 100 150 200 250 300 350 Data / Pred.

0.5 0.75 1 1.25 1.5

Events / 25 GeV 20 40 60 80 100 ATLAS

1

= 13 TeV, 36.1 fb s Dilepton

1 4j ≥

SR Post-Fit Data H t t + light t t 1c ≥ + t t 1b ≥ + t t + V t t t Non-t Total unc. H (norm) t t Pre-Fit Bkgd.

13

SLIDE 14

t¯ tH(b¯ b) uncertainties

Phys. Rev. D 97 (2018) 072016

1c: ISR / FSR ≥ + t t 1b: shower recoil scheme ≥ + t t 5F vs. nominal

HERPA

1c: S ≥ + t t 3b normalization ≥ 1b: tt+ ≥ tt+ H: cross section (QCD scale) t t Jet energy resolution: NP I 0.10 ± 1b) = 1.24 ≥ k(tt+ b-tagging: mis-tag (light) NP I H: PS & hadronization t t 1b: ISR / FSR ≥ + t t 1b: PS & hadronization ≥ + t t 4F vs. nominal

HERPA

1b: S ≥ + t t 5F vs. nominal

HERPA

1b: S ≥ + t t

µ ∆ 1 − 0.5 − 0.5 1

: µ Pre-fit impact on θ ∆ + θ = θ θ ∆

θ

= θ : µ Post-fit impact on θ ∆ + θ = θ θ ∆

θ

= θ

Nuis. Param. Pull

ATLAS

1

= 13 TeV, 36.1 fb s

k(t¯

t+ ≥ 1b) is a normalization parameter

Measure

1.24 ± 0.10 (uncertainty statistical)

The four uncertainties with the largest impact on the limit are all t¯

tb related

A better understanding of t¯

tb¯ b crucial for this channel to be useful

14

SLIDE 15

CMS uncertainties

JHEP 03 (2019) 026

CMS result: µ = 0.72 ± 0.24 ± 0.38 ATLAS result: µ = 0.79 ± 0.29 ± 0.53

15

SLIDE 16

t¯ tb¯ b predictions

Clearly, an improved understanding of t¯

tb¯ b is need for searches/measurements of t¯ tH(b¯ b)

Many other searches have large t¯

tb¯ b backgrounds (SUSY, four top production . . . )

Predicting t¯

tb¯ b is challenging

Massive b-quarks in the matrix element, large scale differences (mt versus

mb)

NLO predictions of t¯

tb¯ b started to arrive about five years ago

Some surprising results . . .

16

SLIDE 17

t¯ tb¯ b predictions

Phys. Lett. B734 (2014) 210
NLO t¯

tb¯ b production with massive b-quarks in the matrix element using the four flavour scheme

The effect of g → b¯

b splitting in the parton shower is important (MC@NLO vs. MC@NLO2b)

The contribution of the right

diagram below is surprisingly large

Parton shower effects still

important at NLO

Cross-section uncertainties range

from 20-40% (depending on fiducial cuts)

17

SLIDE 18

Measuring t¯ tb¯ b

Experimental input is required to move our understanding of t¯

tb¯ b forward

ATLAS has performed a measurement of t¯

t with additional heavy-flavour jets at 13 TeV, using data collected in 2015 & 2016

Fiducial cross-sections measured
We DO NOT attempt to identify which b-jets are from the top quarks and

which are considered “additional”

The measurement therefore includes “QCD” t¯

tb¯ b, t¯ tH and t¯ tZ H Z/γ

18

SLIDE 19

Analysis outline

1. Select an inclusive (≥ 2 b-jets) sample of t¯

t events

2. Categorise simulated t¯

t events based on the ”flavours” of the jets in the

event. Use these to create templates from t¯

t simulation

3. Fit the templates to data in a discriminating variable
4. From the results of this fit, measure inclusive and differential fiducial

cross-sections The analysis is performed in two channels

ℓ+jets
eµ

19

SLIDE 20

Analysis selection

Both channels require the ATLAS detector to be fully operational
A primary vertex with at least two tracks
Single electron/muon triggers with pT > 20(26) GeV for muons and

pT > 24(26) GeV for electrons in 2015 (2016) ℓ+jets

1 ℓ(e/µ) with pT > 27 GeV
≥ 5 jets with pT > 25 GeV,

|η| < 2.5

≥ 2 tagged at the 60%

b-tagging efficiency WP eµ

1 e and 1 µ with pT > 27

GeV

Qe · Qµ = −1
≥ 2 jets with pT > 25 GeV,

|η| < 2.5

≥ 2 jets tagged at the 77%

b-tagging efficiency WP

20

SLIDE 21

Number of b-tags after pre-selection

JHEP 04 (2019) 046

Events

1 10

2

10

3

10

4

10

5

10

6

10

7

10

8

10

9

10

Data t t H t t V t t Single top *+jets γ Z/ Diboson NP & fake lep. Syst.

ATLAS

1

= 13 TeV, 36.1 fb s channel µ e @77% pre-fit b 2 ≥

jets

b

N

2 3 4 ≥

Data/Pred.

0.6 0.8 1 1.2 1.4 1 10

2

10

3

10

4

10

5

10

6

10

7

10

8

10 Events

2 3 4 5 ≥

jets

b

0.8 0.9 1 1.1 1.2

Data/Pred.

Data t t H t t V t t Single top W+jets Z+jets Diboson NP & fake lep. Syst.

ATLAS

1

= 13 TeV, 36.1 fb s lepton+jets channel @60% pre-fit b 2 ≥ j 5 ≥

N

After pre-selection there is a slope in the data / MC ratio in the number of

b-jets distribution

The number of events with ≥ 3 b-jets is under-estimated
We want to identify the cause of this. Is it due to modelling or an

experimental effect (flavour tagging?)?

21

SLIDE 22

Leading b-jet pT

JHEP 04 (2019) 046

We can also look at jet variables
Here we can clearly see the ”purity” of the sample we are dealing with
Non-t¯

t backgrounds are small contributions in both channels

Events / GeV

10 20 30 40 50

Data t t H t t V t t Single top *+jets γ Z/ Diboson NP & fake lep. Syst.

ATLAS

1

= 13 TeV, 36.1 fb s channel µ e @77% pre-fit b 3 ≥

[GeV]

1

b T

p

50 100 150 200 250 300

Data/Pred.

0.6 0.8 1 1.2 1.4 50 100 150 200 250 300 350 400

Events / GeV

50 100 150 200 250 300

[GeV]

1

b T

p

0.6 0.8 1 1.2 1.4

Data/Pred.

Data t t H t t V t t Single top W+jets Z+jets Diboson NP & fake lep. Syst.

ATLAS

1

= 13 TeV, 36.1 fb s lepton+jets channel @60% pre-fit b 3 ≥ j 5 ≥

22

SLIDE 23

Discriminating t¯ tb, t¯ tc and t¯ tl

JHEP 08 (2018) 89

To discriminate between the various cases, we use the output of the

b-tagging variable

The output of the b-tagging algorithm is split into five bins, each of which

is calibrated

The tightest b-tagging working point (5) is 60% efficiency and has a

light(c)-jet rejection rate of ≈ 1550(35)

1 − 0.8 − 0.6 − 0.4 − 0.2 − 0.2 0.4 0.6 0.8 1

MV2c10 BDT output distribution

3 −

10

2 −

10

1 −

10 1

Event fraction

|<2.5 η >20 GeV, |

T

jet p

ATLAS Simulation t = 13 TeV, t s

b-jets c-jets Light-flavour jets 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1

b-jet efficiency

10

2

10

3

10

4

10

Background rejection

|<2.5 η >20 GeV, |

T

jet p

Light-flavour jet rejection c-jet rejection

ATLAS Simulation

23

SLIDE 24

eµ channel

JHEP 04 (2019) 046

In the eµ channel the fit is

performed using the b-tagging discriminant of the jet with the third largest value of DMV2

t¯

tc and t¯ tl templates are combined

A systematic uncertainty is

included by varying the normalisation of the t¯ tc template by ±40% before combining with the t¯ tl template

The best fit value scales the t¯

tb template by ≈ 1.4

Events

500 1000 1500 2000 2500

Data l t c+t t t 0.04 ± = 1.05

cl

α b t t 0.06 ± = 1.37

b

α t Non-t Pre-fit

channel µ e @77% b 3 ≥

ATLAS

1

= 13 TeV, 36.1 fb s

tag discriminant bin

b 3 4 5 Pred./Data

0.6 0.8 1 1.2 1.4

Pre-fit Post-fit

24

SLIDE 25

ℓ+jets channel

JHEP 04 (2019) 046 2

10

3

10

4

10

5

10

6

10

7

10 Events

tag discriminant bin

b

0.8 1 1.2

Pred./Data

jet

rd

3 1 2 3 4 5 2 3 4 5 3 4 5 4 5 5 jet

th

4 1 1 1 1 1 2 2 2 2 3 3 3 4 4 5 Data l t t c t t b t t t Non-t Pre-fit Pre-fit Post-fit

ATLAS

1

= 13 TeV, 36.1 fb s lepton+jets channel @60% b 2 ≥ j 5 ≥ 0.003 ± 0.962 =

l

α 0.06 ± 1.59 =

c

α 0.02 ± 1.11 =

b

α

In the ℓ+jets channel there are

always at least four jets and so the fit is performed using the b-tagging discriminants of the two jets with the third and fourth largest values of DMV2

2D fit flattened to 1D in the figure
t¯

tc and t¯ tl templates are treated separately

The best fit value scales the t¯

tb template by ≈ 1.1

Can see that the final bin in the

distribution, which is equivalent to four very tight b-tags, is very pure in t¯ tb events

25

SLIDE 26

Applying the correction factors

JHEP 04 (2019) 046

Events

1 10

2

10

3

10

4

10

5

10

6

10

7

10

8

10

9

10

Data t t H t t V t t Single top *+jets γ Z/ Diboson NP & fake lep. Syst.

ATLAS

1

= 13 TeV, 36.1 fb s channel µ e @77% post-fit b 2 ≥

jets

b

N

2 3 4 ≥

Data/Pred.

0.6 0.8 1 1.2 1.4 1 10

2

10

3

10

4

10

5

10

6

10

7

10

8

10 Events

2 3 4 5 ≥

jets

b

N

0.8 0.9 1 1.1 1.2

Data/Pred.

Data t t H t t V t t Single top W+jets Z+jets Diboson NP & fake lep. Syst.

ATLAS

1

= 13 TeV, 36.1 fb s lepton+jets channel b 2 ≥ j 5 ≥

We can apply these correction factors back to our poorly modelled

distributions

There is a clear improvement in the agreement between the data and the

prediction

26

SLIDE 27

Applying the correction factors

JHEP 04 (2019) 046

Events / GeV

10 20 30 40 50

Data t t H t t V t t Single top *+jets γ Z/ Diboson NP & fake lep. Syst.

ATLAS

1

= 13 TeV, 36.1 fb s channel µ e @77% post-fit b 3 ≥

[GeV]

1

b T

p

50 100 150 200 250 300

Data/Pred.

0.6 0.8 1 1.2 1.4 50 100 150 200 250 300 350 400

Events / GeV

50 100 150 200 250 300

[GeV]

1

b T

p

0.6 0.8 1 1.2 1.4

Data/Pred.

Data t t H t t V t t Single top W+jets Z+jets Diboson NP & fake lep. Syst.

ATLAS

1

= 13 TeV, 36.1 fb s lepton+jets channel b 3 ≥ j 5 ≥

We can apply these correction factors back to our poorly modelled

distributions

There is a clear improvement in the agreement between the data and the

prediction

27

SLIDE 28

t¯ tb fiducial cross-sections

JHEP 04 (2019) 046

10

1

10

2

10

3

10

4 fid [fb]

e ( 4b) e ( 3b) lepton+jets ( 4b) lepton+jets ( 3b)

Data - ttX(X = H, V)

Stat. uncert.

Total uncert. Sherpa 2.2 ttbb (4FS) Powheg+Pythia8 ttbb (4FS) PowHel+Pythia8 ttbb (5FS) PowHel+Pythia8 ttbb (4FS)

0.5 1.0 1.5 Pred./(Data - ttX) ATLAS

s = 13 TeV, 36.1 fb

1

Using these correction factors we can measure t¯

t + b-jet fiducial cross-sections

Data here refers to the measured cross-section
The t¯

tH and t¯ tZ components are subtracted from the results to allow for easy comparison with QCD t¯ tb¯ b predictions

Measured fiducial cross-sections generally larger than t¯

tb¯ b predictions by ≈ 1σ

Confirms what had been seen in related analyses (t¯

teµ, t¯ tH)

Uncertainties range from 13 – 28 %

28

SLIDE 29

Uncertainties

JHEP 04 (2019) 046

Source Fiducial cross-section phase space eµ ℓ + jets ≥ 3b ≥ 4b ≥ 5j, ≥ 3b ≥ 6j, ≥ 4b

unc. [%]
unc. [%]
unc. [%]
unc. [%]

Data statistics 2.7 9.0 1.7 3.0 Detector+background total syst. 8.5 14 18 12 t¯ t modelling total syst. 10 20 21 12 Total 13 26 28 17

Largest uncertainties due to b-tagging (mistagging light and c-jets) and t¯

t modelling

Improving these areas important to understand t¯

t + b-jets in more detail

29

SLIDE 30

Differential measurements

JHEP 04 (2019) 046

In addition, several differential measurements

are made

Use 3 b-jets in the eµ channel and 4 b-jets in the

ℓ+jets channel

“Simple” variables are chosen
Nb−jets in the eµ channel
The b-jet pTs

pb,1

T , pb,2 T , pb,3 T , pb,4 T

The scalar sum of jet pT and lepton pT (used in

t¯ tH(b¯ b) MVAs) Hjets

T

=

i∈jets

pi

T,

HT = Hjets

T

+ pℓ

T

Properties of the bb system (both the leading

two b-jets and the closest two b-jets) pbb

T , mbb, ∆Rbb

]

1

[GeV

T had

H d

bb t t

σ d

bb t t

σ 1

4 −

10

3 −

10

2 −

10

X t t Data- MC

0.6 0.8 1 1.2 1.4

X t t Data- MC

0.6 0.8 1 1.2 1.4

(4FS) b b t t Powheg+Pythia8 Powheg+Pythia8 (RadHi) Powheg+Pythia8 (RadLo) Powheg+Pythia8

X t t Data- MC

0.6 0.8 1 1.2 1.4

t t Sherpa 2.2 (4FS) b b t t Sherpa 2.2 (5FS) b b t t PowHel+Pythia8 (4FS) b b t t PowHel+Pythia8

[GeV]

T had

H

2

10 × 2

3

10

3

10 × 2

QCD b b t t

σ d

b b t t

σ d

b b t t

σ

QCD b b t t

σ 0.9 1 1.1

MG5_aMC@NLO+Pythia8

ATLAS

lepton+jets channel

1

=13 TeV, 36.1 fb s 4b ≥ 6j, ≥ ) H,V = X ( X t t Data - Powheg+Pythia8 MG5_aMC@NLO+Pythia8 Powheg+Herwig7 Syst. Stat.

The shapes of distributions

are generally reasonably described

30

SLIDE 31

Summary and outlook for t¯ tb¯ b

t¯

t + b-jets measured by ATLAS in two channels

t¯

t + b-jet cross-sections measured to be larger than what is predicted

Results confirm what has been hinted at in other analyses
No major shape differences seen in distributions
We have four times more data on disk than has been analysed
As with other t¯

t measurements, the very clean eµ channel will become the most useful channel

We need to understand which variables in the t¯

tH(b¯ b) MVA cause large modelling uncertainties!

Attempt to assign jets from top decays in next measurement?
MC improvements are on the way!

31

SLIDE 32

The view from the top

The top quark may be a window to

new physics

Testing how it interacts with other

particles, particularly the Higgs, is important

The latest t¯

t measurements from ATLAS are a challenge to theorists!

The future (and present) of top