Dimuon scouting update N. Amin, C. Campagnari, V. Krutelyov, M. - PowerPoint PPT Presentation

Dimuon scouting update N. Amin, C. Campagnari, V. Krutelyov, M. Masciovecchio, U. Sarica, I. Suarez, A. Yagil January 9, 2020

Overview of progress/baseline ⚫ Previous update Baseline selections Exactly 2 OS muons and 1 DV in collections, then ⚫ Reminder of baseline selections → ‣ DV ‣ (x, y, z) errors < (0.05, 0.05, 0.1) cm • Note: took out "Num. muon system hits ‣ chi2/ndof < 5 > 0" because it turns out ‣ 𝜍 < 11 cm ‣ Muon (p T > 3 GeV, | 𝜃 | < 2.4) nMatchedStations is hardcoded to 0 ‣ ID (https://github.com/cms-sw/cmssw/ ‣ Num. tracker layers with meas. > 5 blob/CMSSW_9_2_10/HLTrigger/Muon/ ‣ chi2/ndof < 3 ‣ Isolation src/ ‣ Track isolation < 0.1 HLTScoutingMuonProducer.cc#L161), ‣ 𝛦 R with closest jet > 0.3 ‣ Dimuon kinematics and nValidMuonHits is also always 0, ‣ cos( 𝛦𝜚 (dimuon, DV vector)) > 0 during 2017 data-taking ‣ | 𝛦𝜚 (muon 1, muon 2)| < 2.8 ‣ | 𝛦𝜚 (dimuon, DV vector)| < 0.02 ⚫ In the following slides • Follow-up on 𝛦𝜃 / 𝛦𝜚 variable • Signal information 2018 trigger seeds L1 — OR of • Some 2017 and 2018 data comparisons ‣ DoubleMu4p5_SQ_OS_dR_Max1p2 ‣ DoubleMu0er1p4_SQ_OS_dR_Max1p4 • Synchronization exercise ‣ DoubleMu_15_7 2

Followup on 𝛦𝜃 12 / 𝛦𝜚 12 ⚫ Background tracks from PV are characterized by large values of log 10 (abs( 𝛦𝜃 12 / 𝛦𝜚 12 )) between the two muons ⚫ Proposed a cut to reject events with logabsetaphi>2.0 ⚫ This is pretty independent of signal kinematics (flat ~0.7% e ffi ciency loss), so it does not induce mass-sculpting 10 3

Signals ⚫ We are able to successfully run on your signal in • root://cmseos.fnal.gov//store/user/hroutray/DirectGluonFusion_PhiToMuMu/ ggPhimumu_Phimass2_Phictau0_5_part1 ⚫ If it is useful, we have put some signal with a H → ZdZd → 2 𝜈 +2X model in • env -i gfal-ls -lH root://redirector.t2.ucsd.edu//store/user/namin/ProjectMetis/ HToZdZdTo2Mu2X_params_mzd20_ctau50mm_RAWSIM_v9/ • mzd can be replaced with {2, 5, 8, 10, 15, 20} GeV, ctau with {1, 5, 10, 25, 50} mm (30 points in total) • Generated with 2018 MC configuration 4

2017 vs 2018: triggers ⚫ The DoubleMu4 L1 seed was active/unprescaled for only 75% of 2018 data-taking ⚫ In subsequent slides, use the two separate ORs listed below 2017 trigger seeds 2018 trigger seeds L1 — OR of L1 — OR of ‣ DoubleMu4 _SQ_OS_dR_Max1p2 ‣ DoubleMu4p5 _SQ_OS_dR_Max1p2 ‣ DoubleMu0er1p4_SQ_OS_dR_Max1p4 ‣ DoubleMu0er1p4_SQ_OS_dR_Max1p4 ‣ DoubleMu_15_7 ‣ DoubleMu_15_7 5

2017 vs 2018: kinematics ⚫ Compare key variables for small subsets of data in eras 2017D and 2018C after the baseline selection with relaxed isolation • More plots here leading muon p T subleading muon p T dimuon mass DV 𝜍 log 10 (abs( 𝛦𝜃 12 / 𝛦𝜚 12 )) 6

2017 vs 2018: isolation ⚫ Subleading muon track isolation shown below for all masses (left) and for mass > 5 GeV (right) ⚫ Consistent track isolation quantities between the two years 7

Synchronization exercise ⚫ Propose a synchronization exercise to agree on a selection and have consistency Baseline selections in calculation of a few variables • E.g., what displacement variable to use, and how is it corrected? Exactly 2 OS muons and 1 DV in collections, then ⚫ One file from 2017D: /store/data/Run2017D/ScoutingCaloMuon/RAW/ ‣ DV v1/000/302/033/00000/9C0FCC26-8B8D-E711-8A1F-02163E01273D.root ‣ (x, y, z) errors < (0.05, 0.05, 0.1) cm • Lumis: {"302033": [[1, 39]]} (in golden json, ~5pb -1 ) ‣ chi2/ndof < 5 • ~3.2M events in the file ‣ 𝜍 < 11 cm ( corrected wrt beamspot ) • ~400k events when requiring ≥ 1 DV, ≥ 2 muons) ‣ Muon (p T > 3 GeV, | 𝜃 | < 2.4) • ~7k events with baseline requirements on the right ‣ ID ‣ Num. tracker layers with meas. > 5 ⚫ Make baseline selection on right and dump table of relevant quantities to ‣ chi2/ndof < 3 ‣ Isolation compare. Some details: ‣ Track isolation < 0.1 • DV_rhoCorr : DV 𝜍 corrected wrt beamspot ‣ 𝛦 R with closest jet > 0.3 • lxy : L xy , corrected wrt first PV ‣ Dimuon kinematics • Muon1_pt , Muon1_eta , …: muons p T -ordered ‣ cos( 𝛦𝜚 (dimuon, DV vector)) > 0 • absdphimumu : | 𝛦𝜚 (muon 1 , muon 2 )| ‣ | 𝛦𝜚 (muon 1, muon 2)| < 2.8 • absdphimudv : | 𝛦𝜚 (dimuon, DV vector)| ‣ | 𝛦𝜚 (dimuon, DV vector)| < 0.02 • logabsetaphi : log10(| 𝛦𝜃 12 / 𝛦𝜚 12 |) • minabsdxy : min(|d xy,muon1 |, |d xy,muon2 |) 2017 trigger seeds • excesshits : one of the muons has valid pixel hits > expected pixel hits L1 — OR of ‣ DoubleMu4_SQ_OS_dR_Max1p2 (only if DV_rhoCorr > 3.5) ‣ expected hits computed with code here ‣ DoubleMu0er1p4_SQ_OS_dR_Max1p4 ‣ DoubleMu_15_7 ⚫ Full .csv file with 6700 events located at http://uaf-1.t2.ucsd.edu/~namin/dump/ scouting/synchronization/data_2017D_v1.csv • Small excerpt below event lumi run mass DV_x DV_y DV_rhoCorr lxy Muon1_pt Muon1_eta Muon2_pt Muon2_eta absdphimumu absdphimudv logabsetaphi minabsdxy excesshits 275917 1 302033 1.998553 0.041384 -0.118164 0.095345 0.096442 3.777374 -1.525164 3.693338 -1.935441 0.336688 0.017988 0.085850 0.014101 FALSE 557404 1 302033 3.114660 0.086439 -0.024343 0.008920 0.007603 3.534004 1.285894 3.492911 1.239324 0.914805 0.000662 -1.293223 0.003911 FALSE ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 31731494 39 302033 3.090746 0.144525 -0.168947 0.148740 0.150040 8.232279 1.198552 3.358588 1.584871 0.442521 0.015190 -0.058988 0.021072 FALSE 8 31837797 39 302033 2.872157 0.063578 -0.009160 0.031519 0.030407 5.625279 -1.265867 4.952267 -0.736468 0.086819 0.014454 0.785169 0.001000 FALSE

Summary ⚫ Checked consistency between 2017 and 2018 • Updated baseline selection • Should finalize L1 seeds for the two years ⚫ Synchronization exercise 9

Backup 10

Signal generation/e ffi ciencies ⚫ H(125) → Z D Z D → 2 𝜈 +2X • E ffi ciency of generator-level fiducial denominator requirement with respect to full sample is ~64% • Reco*trigger e ffi ciency below shows e ffi ciency dropping to ~30% due to pixel bias cuts 11

Signal kinematics ⚫ c 𝜐 =50mm, m Z D =20GeV signal MC ⚫ ~12 (38)% of signal has p T <10 (20) GeV 12

Track isolation ⚫ Track isolation computation based on the hltIter2L3MuonMergedNoVtx track collection • Calculated within a cone of 𝛦 R<0.3 for tracks within dz<0.2 and d0<0.1 • Exclude 𝛦 R<0.01 tracks • Exclude leading track if it has p T >2 GeV and 𝛦 R<0.025 ⚫ However, stored ScoutingTrack collection from hltIterL3MuonAndMuonFromL1MergedNoVtx ⚫ Sanity cuts on DV errors eliminate the majority of the peak at 1 max(DV x error, DV y error) DV z error 13

Mass distribution ⚫ Using the baseline selection with mass > 5 GeV, DV 𝜍 > 1 cm, further require • log 10 | 𝛦𝜃 12 / 𝛦𝜚 12 | < 2.0 • no excess pixel hits for either muon • min(|d xy,muon1 |, |d xy,muon2 |) > 0.05 cm ⚫ With respect to (baseline, mass > 5GeV, DV 𝜍 > 1 cm), these 3 selections are ~90% e ffi cient on the signal ⚫ Plot mass distributions separately for baseline selection with isolation requirements, and with relaxed isolation requirements • No data in either case for the 1fb -1 subset of 2018 data • In fact, no data down to DV 𝜍 = 0.3 in isolated case (though low statistics…) require isolation relax isolation 14