1. Measurement of the branching ratios for Standard Model Higgs decays into muon pairs and into Z boson pairs at 1.4 TeV CLIC Gordana Milutinovic-Dumbelovic, I. Bozovic-Jelisavcic, C.Grefe, G.Kacarevic, S.Lukic, M.Pandurovic, P.Roloff, I. Smiljanic On behalf of the CLICdp collaboration 9 th International Physics Conference of the Balkan Physical Union – BPU9, 24-27 August 2015, Istanbul University, Istanbul, Turkey

2. Overview  Introduction to the CLIC project  Higgs production at 1.4 TeV CLIC  Motivation for the measurements  Simulation and reconstruction  H→μ + μ - analysis  H→ZZ * analysis  Conclusions 2 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

3. Introduction to the CLIC project 3 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey CLIC vs. LHC -known initial state -no QCD background CLIC will be a Higgs factory

4. Higgs production at 1.4 TeV CLIC At 1.4 TeV WW-fusion is dominant Higgs production channel. Polarisation can enhance statistics. 4 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

5. Motivation for the measurements 5 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

6. Simulation and reconstruction 6 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

7. H→μ + μ - analysis 7 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey Uncertainty is dominated by limited statistics.

8. Signal and background processes 8 -The cross-sections for all processes with photons in the initial state include both Beamstrahlung and processes with EPA photons. -Signal muon pair is generated in the Higgs mass window (100- 150) GeV. Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

9. Preselection and MVA analysis 9 Preselection efficiency 82% Overall signal efficiency 24% Preselection only Preselection + MVA Two major steps: - Preselection (serves to remove beamstrahlung and background with forward electrons) - MVA (background suppression due to different kinematics)

10. Signal and background PDFs  Fully simulated samples of signal and background are generated to extract PDFs.  PDFs are used to describe pseudo-data (both signal and background). 10 Signal Background Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

11. Toy MC experiments  Pseudo-data are made of: randomly sampled fully simulated signal events and background generated with PDFs  Function f is built from signal and background PDFs to describe pseudo-data  Expected shape of data (signal + background) for each Toy MC is fitted with f to extract the number of signal N s 11 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

12. Statistical uncertainty 12 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

13. Result 13 signal efficiency (ε s )24% 38% Uncertainty is dominated by the limited signal statistics and the irreducible background. Systematic uncertainties (uncertainties of the integrated luminosity and muon identification efficiency, uncertainty on the transverse momentum resolution, uncertainty of the signal count caused by the fit) are negligible compared to the statistical uncertainty. Result is obtained with unpolarised beams. Employment of beam polarisation (-80%) can enhance the Higgs production cross-section by a factor 1.8. Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

14. H→ZZ * analysis 14 BR(H→ZZ * ) ≈ 2.89% ⇒ σ HWW × BR ≈ 7.05 fb BR(Z→qq) ≈ 70 % N s (ZZ * → qqqq ) ≈ 5175/1.5 ab −1 BR(Z→e+e−, Z→µ+µ−, Z →ττ) ≈ 10% N s (ZZ * →qqe+e−, ZZ * →qqµ+µ−, ZZ * → qq ττ ) ≈ 1500/1.5 ab −1 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey Reconstruction of multi jet final states.

15. Signal and background processes 15 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey Beamstrahlung and EPA are considered.

16. Preselection 16 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey qqqq final stateqqll final state -log 10 y 34 <3.5 -log 10 y 23 <3.0 P(b) jet1 <0.95 P(b) jet2 <0.95 find two isolated leptons

17. MVA analysis (4-jet final state) 17 Preselection efficiency 30.2% Overall signal efficiency 18% Preselection only Preselection + MVA

18. MVA analysis (2-jet 2l final state) 18 Preselection efficiency 62% Overall signal efficiency 30.4% Preselection only Preselection + MVA Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

19. Results 19 signal efficiency (ε s ) 18% 18.3% signal efficiency (ε s ) 30% 5.6% For the qqqq final state, uncertainty of the measurement is dominated by background with large x-section and by irreducible background with the same topology as the signal. For the qqll final state, uncertainty of the measurement is also dominated by background with similar signature as the signal. Unpolarised beams are assumed. Uncertainties are statistical only. Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

20. Conclusions 20 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

21. Teşekkürler THANK YOU 21 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

22. 22 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey Most relevant processes Other considered processes Tau decays become relevant if both taus decay into two muons which happens in ~3% of cases. The invariant mass of the di-muon system does not match the Higgs mass window considered in this analysis.

23. Electron tagging 23 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey

24. Lepton identification  We have to identify e - and μ form the qqll final state Track energy > 7 GeV Energy contained in a cone around the track: cos θ<0.995 Impact parameters: d 0 <0.2 mm, z 0 <0.2 mm, R 0 <0.2 mm ECAL/HCAL depositions: 0.025< μ ECAL to HCAL fraction<0.3, e - ECAL to HCAL fraction>0.9 24 leptons Other PFOs Steps to reconstruct a tau: Look for tau ‘seed’ (a high energy, charged track) Add all particles within search cone to seed Check number of charged tracks, isolation, tau mass Initial pT cut for all tracks > 4 GeV pT cut for seed > 10 GeV Impact parameter R 0 : 0.01 - 0.5 Search cone angle < 0.15 rad Isolation energy < 3 GeV Ring particles < 5 Invariant mass < 2. GeV/c 2 37% efficiency in reconstruction of tau pair 87% efficiency in reconstruction of the lepton pair

25. Model independent fit  From individual measurements of σ and σ x BR, couplings are determined via a global fit using:  Model independent fit is only possible at lepton colliders.  All results are limited by σ(HZ) measurement.  The Higgs width is extracted with 5.0% - 3.5 % precision. 25 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey width as a free parameter

26. Model dependent fit  LHC-like constraints: no invisible decays, fixed total width  Sub-percent precision, but strongly depend on fit assumptions  Higgs width extraction with 1.6-0.22% precision 26 Gordana Milutinovic-Dumbelovic BPU9, 24-27 August 2015, Istanbul, Turkey width constrained