1. Djamel BOUMEDIENE, Pascal GAY LPC Clermont-Ferrand
Le boson de Higgs à ILC
Higgs Self coupling
Djamel BOUMEDIENE, Pascal GAY
LPC Clermont-Ferrand

2. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
Outline
1) Introduction
2) lhhh measurement
3) Detector performance impact on lhhh measurement
Energy flow
b-tag choice
c contamination
4) Conclusion
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

3. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
Introduction
Study realized for a center of mass energy of 500GeV
Events generated with Whizard - Pythia
Additional backgrounds w.r.t published analysis
mH = 120GeV, Br(H→bb)=68%
Signal cross section 0.18pb
Dlhhh/lhhh ~ 1.75 DshhZ/shhZ
Presence of 6 jets, 8 jets events → overlap → importance of jet reconstruction (typical final state for ILC physics)
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

4. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
Fast sim choice
Parametric fast simulation used
eflow DE/E = 30%/E
Typical b-tag efficiency = 90% (c contamination 35%)
Test of different geometry of detectors
different eflow
different b-tag
Fast sim allows to vary these parameters
To regenerate all the MC for different geometry detectors is too much difficult
How to simulate different energy flow in the full simulation
Various sets of parameters may lead to the same energy flow resolution
Many parameters :
ECAL
HCAL
ID Part
Confusion …
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

5. Parametric Fast simulation
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

6. Monte Carlo simulation
To obtain a realistic result, the generated luminosity should be greater than expected luminosity (handicap for full simulation)
Processes
σ(pb)
N Generated
Generated luminosity (pb-1)
N expected (L = 500 pb-1)
hhZ
0,18441
15k
81340,49
92,2
Backgrounds
699
1820k
332167 tt
526,4
740k
1880,7
263200
ZZZ
1,051
40k
38059,0
525
tbtb
0,7
20k
28571,4
350 ZZ
45,12
50k
1108,2
22560
nntt
0,141327
141515,8 70
wwz
35,3
130k
3682,7
17650
wtb
16,8
200k
2976,2
8400
eezz
0,287
10k
34843,2
143
nnww
3,627
30k
8271,3
1813
evzw
10,094
60k
5944,1
5047
nnzz
1,08257
18474,6
541
ttZ
0,6975
28673,8
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

7. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
Event reconstruction
Events reconstructed
in 6 jets
in 4 jets
in 2 jets
Jet pairing based on di-jet masses associated to bosons
Jets combined in pairs to test different final states
hhZ, hh(nn), ZZZ, ZZ, WWZ
Signal : 3 Channels
hhqq
6 jets
mh & mZ
hhnn
4 jets
Missing energy Mh
hhll
2 energetic leptons
mZ & mh
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

8. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
b - tagging
Use of the b flavor signature to reduce the background
Global variable (crude approach)
Estimator of # of « b jets » per event taking into account de c and uds contamination (parametrisation of a given vdet)
No b-like jet / event
4 b-like jets / event
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

9. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
Event Preselection
Minimal b flavored content
Event by event basis
Visible energy
2 isolated leptons in a mass window of 25GeV arround mZ
Selection
Evis
b content
2 Isoltd lept
hhqq
>370 GeV
> 3.6 No
hhnn
<370 GeV
>1.8
hhll
Yes
hhll
ZZZ
WWZ
background
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

10. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
Event Selection
Multivariable Method
Global inputs :
Visible energy
Sphericity
Reconstructed inputs :
Number of jets (ycut)
²hhZ ²hh ²2-bosons ²3-bosons : hypothesis based on recontructed masses
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

11. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
NN Output
hhqq
hhll
hhnn
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

12. Expected Number of events L = 2 ab-1
Channel hhqq
Signal : 67.3
Background : 93
tt 9.2
tbtb 37.
ZZZ 16.4
ttZ 18.8
Channel hhnn
Signal : 20
Background : 14.6
ZZ 2.4
tbtb 6.8
Channel hhll
Signal 10.3
Background : 26.8 tt
tbtb
Significance = 6.6 s
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

13. Measurement of the Cross Section and hhh coupling
shhZ is measured using maximum Likelihood method assuming a Poisson law distribution
2 dimensional distribution is fitted :
NN Output X b-tag
Pseudo experiment method is used to evaluate the expected statistical error
For (eb=90% and eflow 30%) : Dlhhh/lhhh = 16%
Dshhh/shhh = 9 %
Channel weights
hhqq 40%
hhnn 25%
hhll %
L=2ab-1)
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

14. Impact of detector performance
Move from a realistic fast sim to a smearing of the visible energy of partons
Simulation of different Energy flow
Information at parton level
Merge all together every daughter (except n’s) of the quark in one object
BUT it cannot be done formally because of string model
Method : clusterisation of the daughters to form the parton direction
These objects are smeared afterwards → simulated jets
An optimistic approach for bad energy flow resolutions
Smearing of parton energy
Different eflow
Range : 0%/E → 130% /E
 perfect energy flow
Jet pairing is changed (based on boson masses)
lhhh measurement
For each detector hypothesis :
New NN trained
Cuts fully optimized
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

15. Dlhhh/lhhh versus DE/E
Direct eflow contribution
Precision loss should be more important for high DE/E (eflow contribution is underestimated)
Perfect clusterisations
Dl/l increases with DE/E
Gain when eflow are added = 22.5% → 16%
This reduction corresponds to a reduction of required luminosity of a factor 2 : without the energy flow 4 ab-1 are needed to reach 16% (optimistic)
This result is for a given b-tag/c-tag and for an un-optimized eb choice
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

16. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
b-tag choice
For a given b-tag/c-tag parametrisation
b-tag efficiency (eb) choice = analysis choice
efficiency ↔ purity (c contamination)
Which eb optimizes lhhh measurement ?
Previous working point :
eb=90%
ec =35% (45GeV jet)
Ejet = 45 GeV
Very pure taging
c-contaminated
Low efficiency
High efficiency
(Richard Hawkings parametrisation)
Tuned for back-to-back jets
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

17. Study of Detector performance
Simulation of different b-tag efficiencies
b-tag is defined statistically (from true jet flavor)
c-tag ↔ b-tag respects the Hawkings parametrisation
Tested range : 40% → 95%
For different eflow
lhhh measurement
For each eflow resolution :
NN trained
Cuts optimized
77 analyses are tested (for each combination eflow x btag)
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

18. Dlhhh/lhhh versus b-tag eff.
eb ~ 65%
~ same for different eflow
For eflow = 30%/E
A precision of 15% is reachable on lhhh with the optimal eb
eflow=120%
eflow=0%
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

19. Which background is sensitive to b-tag choicett
tt → WbWb → bbccss
c-jets tagged as b-jets ↔ bbbbqq like final state ↔ hhZ like final state
bbbbqq is the hhZ flavor signature used in the selection
hhZ
tbtb
ZZZ & ttZ tt
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

20. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
Dlhhh/lhhh versus b-tag & EFlow
Expected statistical error (in %) for a lumi of 2ab-1
For a given c-tag / b-tag parametrisation
ec= 35%
Best eb corresponds to
ec < 5%
euds negligible
What is the dependence Dlhhh/lhhh ↔ b-tag purity for a given efficiency (different b-tag/c-tag) 
15%
ec~ 5%
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

21. Dlhhh/lhhh versus b-tag eff.
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

22. b-tag purity ( c tagging)
Previous working point
ec= 35%
ec= 25%
Tested on hhqq
Using eb = 90%
For a c contamination reduction of 30% (typically from 35% to Ejet=45GeV)
Dlhhh/lhhh is reduced by 12%
Preleminary
hhqq
ec= 25%
ec= 35%
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay

23. D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay
Summary & Outlook
Expected statistical precision on lhhh is evaluated to 15% with a typical detector configuration (L=2ab-1)
Fast simulation approach is useful to define the best detector at first level
Effect of 30%/E EFlow is a factor 2 on luminosity (maybe more dur to the optimistic evaluation at 60%/sqrt(E))
b/c tag is of importance and may be convoluate with Calorimeters
Jets clusterisation is also of importance and not completey treated yet.
Obviously the analysis it self may be improved (e.g. jet charge)
04/05/2007
D. BOUMEDIENE | Higgs self coupling | ILC Soft Orsay