GARLIC and ILDopt Marcel Reinhard LLR Ecole polytechnique

GARLIC and ILDopt Marcel Reinhard LLR – Ecole polytechnique Marcel Reinhard ILD LLR Workshop 02/09, Seoul - Ecole Polytechnique

GARLIC Pre-Clustering via hit distance: Ro. I Remove hits close to extrapolated tracks Per Ro. I: Seed finding: 2 -dim energy projection of 7 X 0 Per Seed: Core building (hits close to seed axis) Neighbor clustering (front to back) Simple verification (min. En/hits, distance track) Gap correction ANN rejection Satellite merging Additional iteration for big clusters Final energy estimation Marcel Reinhard LLR - Ecole Polytechnique 2

Whats new in GARLIC? • Leakage correction: very difficult to parametrise • Pre. Shower hits in projection of the TPC endcap – X 0 = 0. 168 (with projection on TPC electronics) – X 0 = 0. 034 (elsewhere) • Works on all cell sizes including gap correciton – Tested 5 x 5, 10 x 10, 20 x 20 mm 2 Marcel Reinhard LLR - Ecole Polytechnique 3

GARLIC for ILD Marcel Reinhard LLR - Ecole Polytechnique 4

Single particle efficiencies with cuts Works down to 150 Me. V Efficiency: >96% @ 500 Me. V 98. 2% @ 1 Ge. V >99% for E>2 Ge. V Marcel Reinhard LLR - Ecole Polytechnique Pion interactions in tracker create fake clusters that are impossible to reject 5

Efficiencies by angle: 1 Ge. V Flat in Gaps visible in Marcel Reinhard LLR - Ecole Polytechnique 6

Efficiencies by angle: 10 Ge. V Flat in Gaps visible in Marcel Reinhard LLR - Ecole Polytechnique 7

Efficiencies by angle: 50 Ge. V Flat in Barrel-Endcap overlap visible in Marcel Reinhard LLR - Ecole Polytechnique 8

Cluster energy in : 1 Ge. V Marcel Reinhard LLR - Ecole Polytechnique 9

Cluster energy in : 10 Ge. V Marcel Reinhard LLR - Ecole Polytechnique 10

Cluster energy in : 50 Ge. V Impact on resolution? Local correction sufficient? 8 vs. 12 -fold geometry? Marcel Reinhard LLR - Ecole Polytechnique 11

Cluster energy in : 1 Ge. V Marcel Reinhard LLR - Ecole Polytechnique 12

Cluster energy in : 10 Ge. V γ e Marcel Reinhard LLR - Ecole Polytechnique 13

Cluster energy in : 50 Ge. V γ e Marcel Reinhard LLR - Ecole Polytechnique 14

Reduce Barrel-Endcap gap: 40 mm Marcel Reinhard LLR - Ecole Polytechnique 15

_ Cell-size: qq events Marcel Reinhard LLR - Ecole Polytechnique 16

_ Performance: qq at 400 Ge. V Pion interactions not suppressed Reasonable performance Marcel Reinhard LLR - Ecole Polytechnique 17

GARLIC on ’s Marcel Reinhard LLR - Ecole Polytechnique 18

ZH → sim A 1 sim Rec 90. 8 1. 9 0. 6 Rec 1. 1 86. 5 10. 9 a 1 Rec 8. 1 11. 6 88. 5 Marcel Reinhard LLR - Ecole Polytechnique 19

Outlook • • Refine gap-correction Use hit count for low energy Develop better leakage correction Pre. Shower hits negligible for resolution, differences in populations need to be checked 8 or 12 -fold geometry? Need to look at loss in energy resolution How much can the Barrel-Endcap-gap be reduced? Alternatively need an approach to cluster among “ 2 ways” How to deal with interactions? / /a 1 separation for different cell-sizes Marcel Reinhard LLR - Ecole Polytechnique 20