[PPT] - 2HDM interpretations Nick Amin April 1, 2019 Overview Slides PowerPoint Presentation

SLIDE 1

2HDM interpretations

Nick Amin April 1, 2019

SLIDE 2

↑

⚫ Slides looking into using a new MG model file for type2 2HDM from last year

are in backup

⚫ Summary of those slides

I got consistency between ttH xsecs from (1) ATLAS (2) Nate’s model (3)

new MG model

I did not get consistency for tHW, tHq between (2) and (3)

⚫ That’s the starting point for these slides

Technical setup
BR(H/A→tt̅) from LHCHXSWG
Reproducing Nate’s xsecs with the new MG model by decoupling charged

Higgses

⚫ Useful links:

HIG-17-027 / AN-2017/202 discusses H/A→tt̅

Overview

2

SLIDE 3

⚫ Using 2HDMtII_NLO model out of the box with the proc card below

LO, no extra partons
Default (dynamical) MG factorization/renormalization scales, nn23lo1 PDF
Using 5FS via define p = p b b~

⚫ Scan over particle mass, tan(𝛾), sin(𝛾-𝛽) for ttX, tXW, tXq for X=h2 (H), h3 (A) ⚫ Important note:

Cannot decay via "pp > t t~ h2, h2 > t t~" since the widths for h2 and h3 are set to 1.0 by

default.

This means I’m just calculating the production cross-section and can’t decay the h2/h3 in MG, but

from LHCHXSWG ROOT files, BR(H/A→tt̅)~1 when mH/A>2mtop (next slide)

Technical Details

3

import 2HDMtII_NLO define p = p b b~ define j = p define tpm = t t~ define wpm = w+ w- define qpm = u c d s u~ c~ d~ s~ b b~ generate p p > tpm qpm h2

utput temp -nojpeg

launch set run_card ebeam1 6500.0 set run_card ebeam2 6500.0 set run_card nevents 5000 set param_card mass 25 125 # h1 set param_card frblock 1 1.0 # tanbeta set param_card frblock 2 1.0 # sinbma set param_card mass 35 550 # H2 set param_card mass 36 550 # H3 set param_card mass 37 1000

assumptions keep SM higgs at 125GeV tan𝛾=1, sin(𝛾-𝛽)=1 mH=mA=chosen mass mH± set to high values (more on this later)

SLIDE 4

BR(H/A→tt̅)

4

⚫ https://twiki.cern.ch/twiki/bin/view/LHCPhysics/LHCHXSWGMSSMNeutral has the following text

"This scenario yields a SM-like Higgs boson h also for low values of mA at about tanβ=7 due to alignment. In contrast to the

definition in arXiv:1808.07542 the ROOT files start only at mA > 120 GeV due to a theoretically inaccessible region at low mA < 120 GeV and tanβ>10."

⚫ And this ROOT file

mh125_align_13.root: 13 TeV, tanβ= 1-20, mA = 120-1000 GeV
ROOT file does not have values for tan𝛾<1

⚫ Below, plot contours of BR in the tan𝛾-mass plane separately for H and A ⚫ Note the difference in color scale — pseudoscalar A has BR~1 even for tan𝛾 reaching ~2-3, but scalar H BR drops off very close to

tan𝛾=1

⚫ Subsequent plots in these slides will have BRs multiplied in (even though it doesn’t make a visible difference)

SLIDE 5

⚫ Nate’s model does not explicitly have H± particles, while the new MG one

does

By default, mH± is 140GeV

⚫ Try decoupling those particles in the new model in 4 ways

1. set width of H± to 0GeV - doesn’t do much compared to backup s16
2. set mH± to 100TeV
3. set mH± to 10TeV
4. set mH± to 1TeV

H± differences

5

SLIDE 6

⚫ Nate’s model does not explicitly have H± particles, while the new MG one

does

By default, mH± is 140GeV

⚫ Try decoupling those particles in the new model in 4 ways

1. set width of H± to 0GeV - doesn’t do much compared to backup s21
2. set mH± to 100TeV - much better agreement with 2016 xsecs
3. set mH± to 10TeV
4. set mH± to 1TeV

H± differences

6

SLIDE 7

⚫ Nate’s model does not explicitly have H± particles, while the new MG one

does

By default, mH± is 140GeV

⚫ Try decoupling those particles in the new model in 4 ways

1. set width of H± to 0GeV - doesn’t do much compared to backup s21
2. set mH± to 100TeV - much better agreement with 2016 xsecs
3. set mH± to 10TeV - nearly identical to 100TeV
4. set mH± to 1TeV

H± differences

7

SLIDE 8

⚫ Nate’s model does not explicitly have H± particles, while the new MG one

does

By default, mH± is 140GeV

⚫ Try decoupling those particles in the new model in 4 ways

1. set width of H± to 0GeV - doesn’t do much compared to backup s21
2. set mH± to 100TeV - much better agreement with 2016 xsecs
3. set mH± to 10TeV - nearly identical to 100TeV
4. set mH± to 1TeV - gives very large tHW/tAW xsecs

H± differences

8 mind the crazy scale

SLIDE 9

⚫ After solving three remaining issues (?), we can make a tan𝛾 vs mass exclusion plot ⚫ Flipping between s5 and s6 is satisfying — close agreement (<10%) between new

model and Nate’s model

So we have a consistent model that can corroborate results using Nate’s model
1. We already have an interpretation with Nate’s numbers. Do we have to update it

to the latest model? I guess so, otherwise the tan𝛾 vs mass exclusion plot will be using different cross-sections than the 𝜏*BR vs mass?

⚫ The new model shows that not decoupling H± and keeping it at a default value of

140GeV will lead to tHW/tHA having ~0 cross-section

And when we set the mass to 1TeV (slide 8), tHW/tHA cross-sections are very large

(O(pb))

2. Do we continue with high mass H±?
3. BR(H/A→tt̅) values provided by the LHCHXSWG do not go below tan𝛾=1
Though I think it’s safe to assume BR=1 below 1

Intermediate summary

9

SLIDE 10

⚫ Continue by decoupling charged higgses (setting mass to 10TeV) ⚫ Generate cross-sections for {H,A} * {ttX,tWX,tqX} * {mass values} * {tan𝛾 values} ~ 1k points marked

with x on top of cross-section*BR contours

Assume mH=mA
This is just the plot on slide 7 but for many different tan𝛾 values

Continuing

10

scalar pseudoscalar

SLIDE 11 ⚫ Now take upper limits from the exclusion plots we already had (cross-section vs mass for H and A separately) and rescale the

upper limits to these new cross-sections

Note, I need to update the cross-sections and re-run the looper for a more correct answer (since relative mixture of single

top and tt processes can make a difference)

⚫ For each mass value (vertical slice), find upper limit on tan𝛾 by linearly interpolating the cross-section in that slice ⚫ ATLAS expected, observed from 2016 analysis (s14) overlaid in red for ttH — not a fair comparison since our exclusion

curve includes single top as well

Around mH~600, ATLAS expected exclusion is nearly equivalent to ours.
Looking at the blue points in the top right plot of s16, I think it’s because they have larger xsecs for ttH and m600 is a bit

jumpy

Continuing

11

scalar pseudoscalar

SLIDE 12

SLIDE 13

More 2HDM checks

Nick Amin October 27, 2018

SLIDE 14

⚫ Goal is to compare 2HDM results from CMS + ATLAS +N.

Craig's paper results on equal footing

⚫ And also I have some misc plots/dumps

Overview

14

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 [TeV]

H

m

3 −

10

2 −

10

1 −

10 1 ) [pb] t t → BR(H × H) t t → (pp σ

Theory (NNLO): = 0.3 β tan = 0.5 β tan = 1.0 β tan Observed limit Expected limit σ 1 ± σ 2 ± All limits at 95% C.L.

1

= 13 TeV, 36.1 fb s SS dilepton / trilepton + b-jets t t → 2HDM type-II H ATLAS [GeV]

H

m 400 500 600 700 800 900 1000 β tan 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Excluded region Observed Expected σ 1 ± σ 2 ± All limits at 95% C.L.

ATLAS

1

= 13 TeV, 36.1 fb s SS dilepton / trilepton + b-jets t t → 2HDM type-II H

(b)

[GeV]

H/A

m 400 500 600 700 800 900 1000 β tan 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Excluded region Observed Expected σ 1 ± σ 2 ± All limits at 95% C.L.

ATLAS

1

= 13 TeV, 36.1 fb s SS dilepton / trilepton + b-jets t t → 2HDM type-II A/H

(GeV)

H

m

350 400 450 500 550

) (fb) t t → BR(H × ,tW,tq)+H) t (t → (pp σ

20 40 60 80 100 120 140 160

95% CL Observed scalar theory σ experiment σ 2 ± 1 and ± 95% CL Expected

(13 TeV)

1

35.9 fb

CMS

(a) (GeV)

A

m

350 400 450 500 550

) (fb) t t → BR(A × ,tW,tq)+A) t (t → (pp σ

20 40 60 80 100 120 140 160

95% CL Observed pseudoscalar theory σ experiment σ 2 ± 1 and ± 95% CL Expected

(13 TeV)

1

35.9 fb

CMS

(b)

Figure 8: Limits at 95% CL on the production cross section for heavy scalar (a) and pseudoscalar

SLIDE 15

⚫ Using 2HDMtII_NLO model out of the box with the proc card below

Default (dynamical) MG factorization/renormalization scales, nn23lo1 PDF
Using 5FS via define p = p b b~

⚫ Scan over particle mass, tan(𝛾), sin(𝛾-𝛽) for ttX, ttX+1jet, tXW, tXq for X=h2 (H), h3 (A) ⚫ Important note:

Cannot decay via "pp > t t~ h2, h2 > t t~" since the widths for h2 and h3 are

set to 1.0 by default. Without properly recalculating the widths as a function of mass/other parameters, the output cross-sections are meaningless.

This means I’m just calculating the production cross-section and can’t decay the

h2/h3 in MG, but the numbers will then be directly comparable with values for 𝜏(pp→tt̅H/A)×BR(H/A→tt̅).

Technical Details

15

set nb_core 10 set automatic_html_opening False import model 2HDMtII_NLO define tpm = t t~ define wpm = w+ w- define p = p b b~ define j = g u c d s u~ c~ d~ s~ b b~ define qpm = u c d s u~ c~ d~ s~ b b~ generate p p > tpm tpm h2

utput output_scan_v1/thw -nojpeg

launch set param_card mass 25 125 # h1 set param_card frblock 1 scan:[0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0,1.1,1.2,1.4,1.6,1.8,2.0,2.2,2.5,3.0,3.5,4.0] set param_card frblock 2 1.0 # sinbma set param_card mass 35 scan:[350,400,450,500,550,600,650,700,750,800,850,900,1000]

SLIDE 16 ⚫ Take ATLAS cross sections for 2HDM at tan𝛾=0.3,0.5,1.0 assuming sin(𝛾-𝛽)=1

for alignment limit (solid lines on right) and compare against what I get from LO MG (dotted lines of the same color)

This is only considering tt̅H
My calculated ones are ~15-20% lower, though the trend is identical

between the two

For reference, dashed purple line on right is what we used in 2016 for tt̅H

→ agreement with dashed blue line suggests we have been using tan𝛾=1

⚫ Possible differences

ATLAS writes "NNLO" on their plot while I’m using LO
ATLAS could also not be using sin(𝛾-𝛽)=1 exactly
Different PDFs/scales?
Turning a log scale plot into x,y pairs

Reproducing ATLAS xsecs

16

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 [TeV]

H

m

3 −

10

2 −

10

1 −

10 1 ) [pb] t t → BR(H × H) t t → (pp σ

Theory (NNLO): = 0.3 β tan = 0.5 β tan = 1.0 β tan Observed limit Expected limit σ 1 ± σ 2 ± All limits at 95% C.L.

1

= 13 TeV, 36.1 fb s SS dilepton / trilepton + b-jets t t → 2HDM type-II H ATLAS

log linear

SLIDE 17

⚫ As a less rigorous check of what our tan𝛾 was in 2016, make use of some numerology ⚫ Focus on the m(H/A)=400 GeV point only

ATLAS only considers tt̅H/A, while CMS considers the two separately but adds (tt̅+tW+tq)H/A
Fortunately this works out nicely if I use the 2016 cross-sections
ttH+ttA is ~40fb
(tt̅+tW+tq)A is ~40fb
So I should be able to compare the expected exclusion points (★) which correspond to

the same cross-section

(Well, the CMS one is 410GeV, not 400, but it’s close still)
Again, this seems to indicate we had tan𝛾~1 in 2016

Confirming our tan𝛾

17

[GeV]

H/A

m 400 500 600 700 800 900 1000 β tan 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Excluded region Observed Expected σ 1 ± σ 2 ± All limits at 95% C.L.

ATLAS

1

= 13 TeV, 36.1 fb s SS dilepton / trilepton + b-jets t t → 2HDM type-II A/H

550

(GeV)

A

m

350 400 450 500 550

) (fb) t t → BR(A × ,tW,tq)+A) t (t → (pp σ

20 40 60 80 100 120 140 160

95% CL Observed

pseudoscalar theory

σ

experiment

σ 2 ± 1 and ± 95% CL Expected

(13 TeV)

1

35.9 fb

CMS

(b)

cross section for heavy scalar (a) and pseudoscalar

SLIDE 18

⚫ Plot our new calculated cross-sections in solid lines for

ttH, ttA, tHq, tAq, tHW, tAW with their 2016 counterparts

🙃 ttH/A agree within 5-7% ☹ tHq/tAq scale differently ☹ tHW/tAW xsecs differ by an order of magnitude

Other processes

18 log linear

SLIDE 19

⚫ Add another jet to ttH

define j = g u c d s u~ c~ d~ s~ b b~
generate p p > t t~ h2
add process p p > t t~ j h2

☹ xsecs increase by more than a factor of 2 (compare with s4)

Adding an extra parton to ttH

19

SLIDE 20

⚫ How much different are xsecs for slightly

lower sin(𝛾-𝛽)<1 (approximate alignment limit)

For ttH with tan𝛾=1, using sin(𝛾-

𝛽)=0.999/0.99/0.9 gives 10/20/80% lower xsec

Small changes in s𝛾𝛽 can change

xsec by a lot

⚫ ttA xsec independent of sin(𝛾-𝛽), and

since the ttA xsec is a bit larger than ttA, this flattens out the dependence on sin(𝛾-𝛽) a little bit

sin(𝛾-𝛽)<1

20

tan𝛾=1

SLIDE 21

Other stuff

21

SLIDE 22

⚫ Just for reference, plot 𝜏 [pb] vs tan𝛾 for a fixed m(H/A)

mass of 400 GeV

Cross-section vs tan𝛾

22 log linear

SLIDE 23

⚫ 8TeV ATLAS analysis in https://arxiv.org/pdf/1707.06025.pdf uses H/A→tt̅ interference with regular tt̅ to set

2HDM limits

2HDM, alignment limit, mH/mA considered separately and also mH=mA degenerate case
At mA=mH=550GeV, excludes tan𝛾<0.92 (1.1), expected (observed)
The ATLAS SS paper excludes tan𝛾<0.85 (0.6), expected (observed)

⚫ At higher masses (700+), SS has tighter exclusion than tt̅ interference analysis

Comparing with another ATLAS 2HDM exclusion

23

500 550 600 650 700 750 mA [GeV] 0.5 1.0 1.5 2.0 tanβ Obs.

Exp. ± 1σ/2σ

Signal Samples 500 550 600 650 700 750 mH [GeV] 500 550 600 650 700 750 mA = mH [GeV] √s = 8 TeV, 20.3 fb−1, all limits at 95% CL

Figure 3: The 95% CL observed and expected exclusion regions for the type-II 2HDM (µ = 1) considering only a pseudoscalar A (left), only a scalar H (middle), and the mass-degenerate scenario mA = mH (right). Blue points indicate parameter values at which signal samples are produced. Table 3: The 95% CL observed and expected exclusion limits on tan β for a type-II 2HDM in the alignment limit considering only a pseudoscalar A (left), only a scalar H (middle), and the mass-degenerate scenario mA = mH (right). A bar (–) indicates that no value of tan β ≥ 0.4 is excluded.

Mass mA mH mA = mH [GeV] tan β:

bs.

exp.

bs.

exp.

bs.

exp. 500 < 1.00 < 1.16 < 1.00 < 0.77 < 1.55 < 1.50 550 < 0.69 < 0.79 < 0.72 < 0.52 < 1.10 < 0.92 600 – < 0.59 < 0.73 – < 1.09 < 0.93 650 – – – – – < 0.62

SLIDE 24

⚫ Not really relevant to 2HDM stuff, but I saw the plots in Nathaniel’s paper comparing

BDT and cut-based analyses for ttH→tttt

⚫ Lowest mass is 0.5TeV, close to SM tttt ⚫ Solid orange line (BDT) is at 1.05fb and dashed line (cut-based) is at 1.15 (~10% better)

Grows to 0.19fb vs 0.38fb (~50% better) at mass of 2TeV where presumably MET/

HT/Boostedness helps a lot

BDT vs cut-based

24

0.5 0.75 1 1.5 2 0.1 0.3 1 2 5

mA [TeV ] (ppH/Atttttt) [fb]

(a) pp → t¯ tH(A) → t¯ tt¯ t

0.5 0.75 1 1.5 2 0.1 0.3 1 2 5

mA [TeV ] (ppH/AtW±tttW±) [fb]

(b) pp → tWH(A) → tW ±t¯ t

Figure 10: (a) Model independent exclusion (orange) and discovery (green) limits at the 14 TeV LHC in the four-top channel. (b) Exclusion (orange) and discovery (green) limits in the three-top

channel. The dashed limits are derived with the cut based analysis presented in Section 4.2 while

the solid limits are derived with the BDT analysis presented in Section 4.3.