Study of Higgsinvisible using kinematic fit method applied jet - PowerPoint PPT Presentation

Study of Higgs→invisible using kinematic fit method applied jet energy resolution of ILD Yu Kato, J.Tian, T.Tanabe, S.Yamashita The Univ. of Tokyo The 55 th General Meeting of ILC Physics Subgroup Feb. 3, 2018

2018/2/3 Outline Study of Higgs->invisible using kinematic fit 2 p Motivation p Idea for improvement p Flow of study u Evaluate jet energy resolution u kinematic fit u Analysis Higgs→invisible p Summary & Plans

0.95% : 0.69% ”Search for Invisible Higgs Decays at the ILC” LCWS2014@Belgrade q Z ν ν ν ν X X q Z q 3 Study of Higgs->invisible using kinematic fit 2018/2/3 Motivation Previous study (A.Ishikawa) (95% CL, 250fb -1 ) it signifies new physics beyond SM (BSM) left pol. : right pol. q l In SM, Higgs decays invisibly through H → ZZ ∗ → 4𝜉 (BR(H → 𝑗𝑜𝑤.)~0.1%) l If BR(H → 𝑗𝑜𝑤.) exceeds SM prediction , l We estimate SM upper limit of BR(H → 𝑗𝑜𝑤.) Dark Matter… SUSY… l Compare result between left & right polarization invisible BSM invisible 𝐶𝑆 H → ZZ ∗ → 4𝜉 ~0.1% 𝐶𝑆 H → XX ~? ? ? % visible visible Ø A. Ishikawa (Tohoku Univ.),

Idea for improvement 2018/2/3 Study of Higgs->invisible using kinematic fit 4 Improve analysis performance kinematic fit apply jet energy resolution method

Flow of study Evaluate jet energy resolution ILD model : ILD_l(s)5_v02 evaluate jet angle resolution also → apply to kinematic fit kinematic fit fit variables : constraint : use MarlinKinfit - fitter engine : OPALFitter apply jet resolution Improve analysis performance [ BSM search using Higgs→invisible ] 2018/2/3 Study of Higgs->invisible using kinematic fit 5 Ø jet energy ＆ cosθ dependence Ø check effect & accuracy of fit

2018/1/17 ILC実験におけるジェットエネルギー分解能評価 及び kinematic fit 手法の研究 6 Evaluate jet energy resolution

Evaluate JER 10k 300 350 400 500 l5 [events] 10k 10k 10k 10k 10k 10k 10k 10k 9k 9k 200 10k s5 [events] 10k 10k 10k 10k 9k 10k 10k 9k 10k 10k 10k 10k use jet clustering: Durham 240 160 120 91 60 40 30 √s [GeV] Setting of Evaluation JER 7 Study of Higgs->invisible using kinematic fit 2018/2/3 � l ILCSoft : v01-19-05 (gcc49) l ILDConfig : v01-19-05-p01 l ILD models : ILD_l5_o1_v02, (ILD_s5_o1_v02) l samples : Z→uds (w/o overlay) l jet resolution definition ◦ use RMS 90 method ◦ Energy 𝐹 = RMS90 𝐹 RMS90 𝐹 𝜏 9 ? ?? = 2 𝑛𝑓𝑏𝑜 CD 𝐹 𝑛𝑓𝑏𝑜 CD 𝐹 ? ?? (J. S. Marshall and M. A. Thomson, ”Pandora Particle Flow Algorithm”, arXiv:1308.4537 [physics.ins-det]) ◦ Angle 𝜀𝜚 = RMS90(𝜚 IJK − 𝜚 MK ) 𝜀𝜄 = RMS90(𝜄 IJK − 𝜄 MK )

ILD model Detailed Baseline Design 2018/2/3 Study of Higgs->invisible using kinematic fit 8 Endcap Barrel θ ILD_l5_v02 ILD_s5_v02 Evaluate JER

Result：Energy dependence 2018/2/3 Evaluate JER 9 Study of Higgs->invisible using kinematic fit sv01-19-05.mILD_l5_o1_v02_nobg 7 ) [%] σ /E = 3.5% E σ /E = 30%/ E E j 6 Overall : (E 31.3/ E -1.97 +0.200 E 90 j j ) / Mean θ Barrel : |cos | < 0.7 5 28.9/ E -1.91 +0.195 E j j θ ≥ Endcap : |cos | 0.7 33.6/ E -1.66 +0.184 E 4 j j j (E 90 RMS 3 0 50 100 150 200 250 E [GeV] j

Result : energy & angle dependence 2018/2/3 apply this result to kinematic fit Evaluate JER 10 Study of Higgs->invisible using kinematic fit sv01-19-05.mILD_l5_o1_v02_nobg 15 15GeV ) [%] 20GeV 30GeV j (E 45.5GeV 90 10 60GeV ) / Mean 80GeV 100GeV 120GeV 5 150GeV j (E 175GeV 90 200GeV RMS 250GeV 0 0 0.2 0.4 0.6 0.8 1 θ |cos |

apply this result to kinematic fit Angular resolution use jet clustering. For evaluation of angular resolution, Evaluate JER Durham algorithm azimuth angle polar angle 11 Study of Higgs->invisible using kinematic fit 2018/2/3 𝜀𝜄 = 𝑆𝑁𝑇 CD (𝜄 IJK − 𝜄 MK ) 𝜀𝜚 = 𝑆𝑁𝑇 CD 𝜚 IJK − 𝜚 MK sv01-19-05.mILD_l5_o1_v02_nobg sv01-19-05.mILD_l5_o1_v02_nobg 0.3 0.08 15GeV 15GeV MC MC 20GeV 20GeV φ θ 0.25 - 30GeV - 30GeV 0.06 REC 45.5GeV REC 45.5GeV 0.2 60GeV 60GeV φ θ 80GeV 80GeV = = 0.15 0.04 100GeV 100GeV φ θ δ 120GeV δ 120GeV 0.1 150GeV 150GeV 175GeV 0.02 175GeV 0.05 200GeV 200GeV 250GeV 250GeV 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 θ θ |cos | |cos |

2018/1/17 ILC実験におけるジェットエネルギー分解能評価 及び kinematic fit 手法の研究 12 kinematic fit

Principle of kinematic fit 2018/2/3 Study of Higgs->invisible using kinematic fit 13 seek minimum of under kinematic constraints method of Lagrange multipliers d.o.f.： kinematic fit

MarlinKinfit : OPALFitter 2018/2/3 Study of Higgs->invisible using kinematic fit 14 For iterative solution : Taylor-expansion of the constraints Convergence condition or kinematic fit ü 𝜀𝜓 R < 0.01% ∩ 𝜀𝐺 V < 10 WX V < 10 WR Y 𝜓 R ∩ 𝐺 [ < 10 W\ ∩ 𝜀 𝜃, 𝜊, 𝜇 < 10 W\ ü all 𝑔

ZH processor 15 invisible X X q q kinematic fit Study of Higgs->invisible using kinematic fit 2018/2/3 sv01-19-05.mILD_l5_o1_v02_nobg 15 15GeV ) [%] 20GeV 30GeV j (E 45.5GeV 90 10 60GeV ) / Mean 80GeV 100GeV 120GeV 5 150GeV j (E 175GeV 90 200GeV RMS 250GeV 0 0 0.2 0.4 0.6 0.8 1 !" H → XX ~???% θ |cos | sv01-19-05.mILD_l5_o1_v02_nobg 0.08 15GeV MC p Fit variables 20GeV θ - 30GeV 0.06 REC 45.5GeV 60GeV θ 80GeV = 0.04 100GeV θ p Z mass constraint δ 120GeV 150GeV 0.02 175GeV 200GeV 250GeV 0 0 0.2 0.4 0.6 0.8 1 p jet mass constraint θ |cos | sv01-19-05.mILD_l5_o1_v02_nobg 0.3 15GeV MC 20GeV φ 0.25 - 30GeV p Implement of jet resolution REC 45.5GeV 0.2 60GeV φ 80GeV = 0.15 100GeV φ δ 120GeV 0.1 150GeV p degrees of freedom 175GeV 0.05 200GeV 250GeV 0 0 0.2 0.4 0.6 0.8 1 θ |cos |

Result：accuracy of fit fit probability Mean > Ndof Ndof ：１ Mean：14.5 a possibility of underestimating parameter error →normal distributed between 0 and 1 fit with well-estimated errors ←peak around０ 2018/2/3 kinematic fit 16 Study of Higgs->invisible using kinematic fit sv01-19-05.mILD_o1_v05.eL.pR sv01-19-05.mILD_o1_v05.eL.pR Events / 2.00 4 OPALFitter 10 fit success : 99.85 % mean = 14.453 sigma = 46.970 3 10 2 10 sv01-19-05.mILD_o1_v05.eL.pR sv01-19-05.mILD_o1_v05.eL.pR 10 Events / 0.01 4 10 OPALFitter fit success : 99.85 % mean = 0.278 1 sigma = 0.313 0 500 1000 1500 2000 χ 2 3 10 χ 2 distribution 2 10 0 0.2 0.4 0.6 0.8 1 Fit Probability

Result：Recoil mass 2018/2/3 improve recoil mass resolution ~20% ↓ISR effect kinematic fit Study of Higgs->invisible using kinematic fit 17 sv01-19-05.mILD_o1_v05.eL.pR sv01-19-05.mILD_o1_v05.eL.pR OPALFitter success : 99.85 % Events / 0.50 GeV 4000 MC: mode = 125.2 sigma = 6.379 3000 2000 1000 0 100 110 120 130 140 150 160 Recoil Mass [GeV] sv01-19-05.mILD_o1_v05.eL.pR sv01-19-05.mILD_o1_v05.eL.pR sv01-19-05.mILD_o1_v05.eL.pR OPALFitter success : 99.85 % OPALFitter success : 99.85 % Events / 0.50 GeV Events / 0.01 before fit: before fit: 1200 600 mean = 130.1 mean = 8.4e-03 sigma = 12.076 sigma = 8.8e-02 1000 after fit: after fit: 800 400 mean = 129.0 mean = -3.3e-04 sigma = 10.496 sigma = 6.9e-02 600 400 200 200 0 0 − − 100 110 120 130 140 150 160 1 0.5 0 0.5 1 Recoil Mass [GeV] Recoil Mass Relative Error

Problems : Z mass distribution 2018/2/3 Error??? kinematic fit Study of Higgs->invisible using kinematic fit 18 sv01-19-05.mILD_o1_v05.eL.pR OPALFitter success : 99.85 % Events / 0.50 GeV MC: before fit: 4 10 mean = 90.9 mean = 90.7 sigma = 5.338 sigma = 10.091 3 10 after fit: mean = 91.3 2 10 sigma = 1.271 10 1 70 80 90 100 110 120 M [GeV] Z

the Cause: 2018/2/3 NewtonFitter OPALFitter Approximate calculation of constraint in OPALFitter kinematic fit 19 Study of Higgs→invisible using kin-fit applied JER of ILD sv01-19-05.mILD_o1_v05.eL.pR sv01-19-05.mILD_o1_v05.eL.pR OPALFitter success : 99.85 % NewtonFitter success : 99.35 % Events / 0.50 GeV Events / 0.50 GeV MC: before fit: MC: before fit: 4 4 10 10 mean = 90.9 mean = 90.4 mean = 90.9 mean = 90.7 sigma = 5.486 sigma = 9.516 sigma = 5.338 sigma = 10.091 3 3 10 10 after fit: after fit: mean = 91.3 mean = 91.2 2 2 10 10 sigma = 1.271 sigma = 0.077 10 10 1 1 70 80 90 100 110 120 70 80 90 100 110 120 M [GeV] M [GeV] Z Z

2018/1/17 ILC実験におけるジェットエネルギー分解能評価 及び kinematic fit 手法の研究 20 Search for BSM using H→invisible