Minimum bias studies etc…. Craig Buttar University of Glasgow Cosners House Forum on LHC startup April 07 1 Outline • LHC start up • Minimum bias events – Why study them? – What are they? • Triggering minimum bias – Minimum bias trigger scintilators (MBTS) – Random trigger and track trigger • Tracking – Reconstructing low pT tracks – Vertex reconstruction – Using pixel tracklets • Underlying event – What is it and why measure it – Measurement with ATLAS • Summary Cosners House Forum on LHC startup April 07 2 LHC Pilot run: 2007 Hardware Commissioning 450GeV Installation Hardware Commissioning Machine checkout 450GeV Beam commissioning 450GeV Engineering Run 450GeV Collisions 450GeV Ramp commissioning Kb 43 43 156 156 ib (1010) 2 4 4 10 * 1 5 (m) 11 11 11 11 intensity per beam 8.6 1011 1.7 1012 6.2 1012 1.6 1013 beam energy (MJ) .06 .12 .45 1.1 2 1028 7 1028 2.6 1029 1.6 1030 event rate 1 1 5 (kHz) 0.7 2.8 10 65 W rate 2 1 5 (per 24h) 0.8 3 11 70 Z rate 3 1 5 (per 24h) 0.08 0.3 1.1 7 Luminosity 1 5 At L=1029, can get ~10M MB events in 1 day with 30% eff Cosners House Forum on LHC startup April 07 Shutdown Events/xing <<1 S. Myers 3 Possible Beam Parameter Evolution at 7TeV/beam Parameters Beam levels Rates in 1 and 5 Rates in 2 (and 8) kb N * 1,5 (m) Ibeam proton Ebeam (MJ) Luminosity (cm-2s-1) Events/ crossing Luminosity (cm-2s-1) Events/ crossing 43 4 1010 11 1.7 1012 2 1.1 1030 << 1 1.2 1030 0.15 43 4 1010 2 1.7 1012 2 6.1 1030 0.76 1.2 1030 0.15 156 4 1010 2 6.2 1012 7 2.2 1031 0.76 4.4 1030 0.15 936 4 1010 11 3.7 1013 42 2.4 1031 << 1 2.6 1031 0.15 936 4 1010 2 3.7 1013 42 1.3 1032 0.73 2.6 1031 0.15 2808 4 1010 2 1.1 1014 126 3.8 1032 0.72 7.9 1031 0.15 2808 5 1010 2 1.4 1014 157 5.9 1032 1.1 1.2 1032 0.24 2808 5 1010 1 1.4 1014 157 1.1 1033 2.1 1.2 1032 0.24 S. Myers Cosners House Forum on LHC startup April 07 4 Why measure min bias? Not exactly what the LHC was built for! But….. • Physics: measure dN/dh|h=0 – Compare to NSD data from SppS and Tevatron • MB samples for pile-up studies – Calorimeter – Physics analyses – Benchmark for sLHC • Overlap with underlying events – analyses eg VBF, Jets… • Demonstrate that ATLAS is operational • Inter-calibration of detector elements – Uniform events • Alignment • Baseline for heavy ions Cosners House Forum on LHC startup April 07 5 inelastic events What are minimum bias events? pp collisions at √s = 14TeV PYTHIA6.323 PHOJET1.12 σtot 101.5 mb 119.1 mb σelas 22.2 mb 34.5 mb σsingle dif (AB→XB) 7.2 mb 5.5 mb σ single dif (AB→AX) 7.2 mb 5.5 mb σdouble dif 10.3 mb 4.1 mb σnon-dif 54.7 mb 69.5 mb Trigger bias: Typically select non-single diffractive (NSD) events Cosners House Forum on LHC startup April 07 6 What are minimum bias events? • Minimum bias are inelastic collisions of two protons – Includes very rare high-pt scatters and very common low-pt scatters • Minimum bias is an experimental definition – Defined by experimental trigger and analysis • Relation between experiment and physics is: measured f f f sd sd dd dd nd inelastic nd inelastic fi are acceptances for different physics processes determined by the trigger Cosners House Forum on LHC startup April 07 7 Event characteristics • Event characteristics – Non-single diffractive~non-diffractive inelastic – Soft tracks pTpeak~250MeV – Approx flat distribution in h to |h|~3 and in f – Nch~30; |h|<2.5 • Trigger rates ~70mb (NSD!) – R~700kHz @ L=1031cm-2s-1 M. Leyton Cosners House Forum on LHC startup April 07 8 Triggering minimum bias • Use forward scintillators (MBTS) • Random trigger – inefficient where there are <<1 events/xing – Validate at L2/EF to reduce inefficiency • Other triggers: jet, lepton triggers – High pt thresholds bias Cosners House Forum on LHC startup April 07 9 MBTS • Trigger scintillation counters mounted on end of LAr calorimeter covering same radii as ID η=2.0 η=3.8 Beam-pipe +η pannel interaction point MBTS UA5 • • Cosners House Forum on LHC startup April 07 To compare to UA5 and CDF data need to understand composition of the sample trigger bias Currently generate inelastic and diffractive parts using PYTHIA – Need to investigate other simulations10 PHOJET Noise and Accidental Suppression – Lab tests • • Eric Feng Noise spectrum fits well to a Gaussian with σ=2.52pC Beyond 9pC (almost 4σ), nonGaussian behavior is possible Accidental rate must be suppressed to ~Hz, limited by EF output-rate of 100Hz suppression required of 10-6-10-7 at 900GeV and 14TeV Cosners House Forum on LHC startup April 07 11 Trigger Efficiencies and Bias • Shown below for each coincidence logic: - Required counter threshold to suppress accidental rate to 1Hz - Corresponding trigger efficiency for each type of event Coincidence Logic Counter Threshold [pC] Trigger Efficiency 900 GeV 14 TeV Non-Diff Single/ Double Diff. Non-Diff Single/ Double Diff. 1+1 10.75 98% 35% 99% 42% 2+2 7.75 97% 27% 98% 36% 3+3 6.25 94% 20% 97% 30% 4+4 5.25 90% 15% 95% 24% 5+5 4.50 85% 10% 92% 19% 6+6 4.00 79% 7% 89% 14% Cosners House Forum on LHC startup April 07 Eric Feng 12 13 MHz eventrate Track Trigger random trigger L2 track trigger to validate random trigger when <<1 events/xing LVL 1 1 kHz eventrate pass-through LVL 2 reduction by a factor 10 Event-Filter 100 Hz eventrate R. Kwee Cosners House Forum on LHC startup April 07 storage 13 Selecting MinBias Events • At L2 can use – no. of spacepoints from pixel and SCT little bias – no. of reconstructed tracks, needed to reject beamgas, halos pt-track > ~1GeV biased • Rejection of empty events, compare to cosmic runs noise spacepoints from ¼ sct-barrel from cosmics data <sct-spacepoints>=3 simulation of noise in whole SCT <sct-spacepoints>=0 R. Kwee Cosners House Forum on LHC startup April 07 14 Rejecting beam-gas Diffractive Beamgas SCT-cut at 20 sp pixel-cut at 900 sp minbias 99.2 86.3 noise 0 0.4 beamgas 77.3 31.6 Diffractive (SD+DD) 50.2 15.0 Cosners House Forum on LHC startup April 07 To reject beam gas need track reconstruction R. Kwee 15 CMS MB trigger: Trigger Towers ORCA Simulation • Use Forward HCAL – Towers: N vs P • Strongly correlated • Trigger should cut on nTowers hit on “N” side and “P” side – e.g. N>10 and P>10 Cosners House Forum on LHC startup April 07 16 Tracking in MB events 1000 events dNch/dh • Acceptance limited in rapidity and pt • Rapidity coverage – Tracking covers |h|<2.5 • pT problem – Need to extrapolate by ~x2 Need to understand low pt charge track reconstruction h Black = Generated (Pythia6.2) Blue = TrkTrack: iPatRec Red = TrkTrack: xKalman dNch/dpT Reconstruct tracks with: 1) pT>500MeV 2) |d0| < 1mm 3) # B-layer hits >= 1 4) # precision hits >= 8 Cosners House Forum on LHC startup April 07 17 pT (MeV) Where is the Momentum Limit ? • Tracker is in principle sensitive to soft tracks 50MeV – Pt = 400 MeV - tracks reach end of TRT – Pt = 150 MeV - tracks reach last SCT layer – Pt = 50 MeV - tracks reach all Pixel layers A.Salzburger 400MeV Cosners House Forum on LHC startup April 07 150MeV 18 Track reconstruction efficiencies • MC/Track cuts: η < 2.5, z0 < 150 mm, d0 < 1 mm • Efficiency: fraction of MC tracks with > 50% matched hits Tracking efficiency >20-50% for pt>200MeV • Pt=200MeV pT (MeV) Cosners House Forum on LHC startup April 07 M.Leyton 19 Track Reconstruction efficiencies by track insertion Dan Clemens efficiency by track insertion reconstruction with New Tracking •Insert a track into a minimum bias event by overlapping a simulated single track event. •Reconstruct merged event and look for inserted single track to determine efficiency. Cosners House Forum on LHC startup April 07 Plot by M.Leyton for comparison : truth matched 20 Minbias Vertex Reconstruction How well can we reconstruct primary vertex position of a minbias event? – A good vertex reconstruction efficiency in multi-pp bunch crossings may allow measurement of σNSD. – Knowledge of primary vertex position can improve signal to noise, eg, H →γγ (UE to Higgs similar to MB) • • Use VxBilloirTools package to fit vertex to minbias tracks. Shown here are residuals (difference between reconstructed and truth) X Vertex Residuasl • Track multiplicity Typically ~250μm precision Cosners House Forum on LHC startup April 07 I. Dawson, K. Prokofiev 21 Tracklet Method Peter Steinberg, BNL η Hits in Layer 1 Exploring methods with sensitivity at lower pT with even fewer pixel-hits Φ Hits in B-layer Primary vertex (standard tracking) Each hit assigned η,Φ relative to primary vertex Calculate Δη,ΔΦ for each pair “Tracklet” defined (now) as Δη<0.01,ΔΦ<0.15 Cosners House Forum on LHC startup April 07 22 Peter Steinberg, BNL Single Event Φ η Hits in B-layer and L1. Black circles are Truth-primary charged hadrons. Clustered pairs of hits visible near truth hadrons. Cosners House Forum on LHC startup April 07 23 Peter Steinberg, BNL “Momentum Reconstruction” pT = 0.011/ΔΦ → Poor person’s tracking code. spatial resolution kicks in above ~2 GeV Cosners House Forum on LHC startup April 07 24 Momentum reconstruction Peter Steinberg, BNL Charged primary hadrons Reconstructed at low pT (down to 100 MeV) the recontructed and truth pT match well, at higher pT reconstruction gets distorted Cosners House Forum on LHC startup April 07 25 Peter Steinberg, BNL Inclusive dNch/dη Charged primary hadrons Reconstructed Current parameters good for η~0, large backgrounds at |η|~2 Cosners House Forum on LHC startup April 07 Focus on |η|<1, look at multiplicity distributions linear response,good resolution 26 CMS Tracking Similar to ATLAS pixel tracklets Cosners House Forum on LHC startup April 07 27 What is the underlying event Hard and soft physics • ISR • beam remnants • central particle production • colour flow/string drawing High PT scatter Beam remnants ISR Cosners House Forum on LHC startup April 07 28 VBF Signal (HWWll) •forward tagging jets •correlated isolated leptons • low hadronic activity in central region •central Higgs production Tagging jet W Z/W Z/W H W Tagging jet Important discovery channel For Higgs in mass range 120-200GeV Cosners House Forum on LHC startup April 07 29 Model Parameter Simple MSTP(82)=1 PARP(81)=1.9 Complex MSTP(82)=4 PARP(82)=1.9 Tuned MSTP(82)=4 PARP(82)=1.8 PARP(84)=0.5 Model CJV Simple (eff) LEPACC (eff) All Vetoes (eff) 0.943 ± 0.003 0.610 ±0.001 0.084 ±0.001 Complex 0.885 ± 0.003 0.575 ±0.001 0.076 ±0.001 Tuned 0.915 ± 0.003 0.589 ±0.001 0.080 ±0.001 Uncertainty at the level of ~6% on CJV Pt>20GeV and ~3% on lepton Giving a total uncertaintly in the range ~8% Cosners House Forum on LHC startup April 07 30 Describing the region transverse to the leading jet Average multiplicity of charged particles in the underlying event associated to a leading jet with Ptljet (GeV). Cosners House Forum on LHC startup April 07 Average pTsum (GeV) of charged particles in the underlying event associated to a leading jet with Ptljet (GeV). 31 Min-Max analysis “Leading Jet” Jet #1 Direction f “Toward” “TransMAX” Jet #1 Direction f “TransMIN” “Away” “Toward” “TransMAX” “TransMIN” “Away” “Back-to-Back” Jet #2 Direction •TransMAX and transMIN sensitive to radiation and soft UE respectively •Back-to-back sample suppresses radiation • difference between tranMAX region and transMIN in leading jet and b-2-b jet sample Cosners House Forum on LHC startup April 07 32 Measuring the underlying event at ATLAS ET jet #1 = 1.63TeV ET jet #2 = 1.16TeV Cosners House Forum on LHC startup April 07 33 33 First physics studies with the underlying event ATLAS - √s=900 GeV ATLAS - √s=14 TeV Reconstructed tracks √s=14TeV Multiplicity of charged particles with pT > 0.5 GeV and |h|< 1 in region transverse to leading jet √s=900GeV ~ 15 days of data taking in 2007 enough to cover up to pT(leading jet) ~ 40 GeV Cosners House Forum on LHC startup April 07 ~ 15 days of data taking in 2007 enough to cover up to pT(leading jet) ~ 40 GeV 34 34 Measuring the underlying event in TeV jets Selecting the underlying event: charged particles (generator level) Njets > 1, |ηjet| < 2.5, ETjet >10 GeV, |ηtrack | < 2.5, pTtrack > 1.0 GeV/c charged particles (generator level) J6 J6 J5 J5 J4 J4 J1 J2 J3 J1 Cosners House Forum on LHC startup April 07 J2 J3 35 Selecting the underlying event: Cosners House Forum on LHC startup April 07 Njets > 1, |ηjet| < 2.5, ETjet >10 GeV, |ηtrack | < 2.5, pTtrack > 1.0 GeV/c 36 summary • Minimum bias and underlying event are uncertain at LHC energies – Energy extrapolation for SppS and CDF not known – dN/dh|h=0 measurement can be done in early running • Triggering – Both ATLAS and CMS are developing minimum bias triggers – Need to understand trigger bias to compare to UA5,CDF…. • Tracking – Low pt tracking is challenging but progress by both ATLAS+CMS • Underlying event can be measured with early data – Need to ensure overlap between jet and mb samples • Waiting for data Cosners House Forum on LHC startup April 07 37 Extra slides Cosners House Forum on LHC startup April 07 38 Day zero start-up looks promising! We are taking REAL DATA ! • SR1 cosmics test stand – – – Final TRT barrel Final SCT barrel 1 sector cabled for both Cosners House Forum on LHC startup April 07 39 First Results • Offline reconstruction worked with little problems within hours – • Debug cabling… First results are very positive – – – See residuals and pulls No alignment or detailed calibration yet ! Reflects excellent mechanical precision… • More for Ringberg Workshop in 2 weeks • I am definitely very optimistic for first data taking having seen this ! – • Detector is on a good way ! Pixels are recovering from cooling problem delay – Finish 3 layer system in time Cosners House Forum on LHC startup April 07 40 Tracking: Startup-Initial Alignment • Very first alignment will be based on: – – – – Mechanical precision Detailed survey data Cosmics data (SR1/Pit) Minimum bias events and inclusive bb • Studies indicate good efficiencies after initial alignment – – Cosners House Forum on LHC startup April 07 Example taken from T.Golling Precision will need Zs and resonances to fix energy scales, constrain twists, etc… M.Elsing Talk in SM meeting 41 Coincidences in MBTS PYTHIA charged-particle level Non-diffractive+SD+DD Diffractive events Non-diffractive only Cosners House Forum on LHC startup April 07 42 MBTS response Non-diffractive inelastic events (sample 005001) Single and double diffractive inelastic events (sample 005002) Single diffractive events: energy is deposited on either side Non-diffractive inelastic events: energy is deposited on both sides (typically) Cosners House Forum on LHC startup April 07 A.Moraes Double diffractive events 43 How many events? • For dN/dh require ~10k • For UE need ~20M MB events to get some with leading jets PT~30GeV • Need to overlap with jet triggers – Prescaled PT>25GeV • ~40M events gives ~10 hits/strip in barrel SCT • CMS quote ~18M to intercalibrate their calorimeter with MB Cosners House Forum on LHC startup April 07 44 Performance goals Cosners House Forum on LHC startup April 07 45 Machine considerations Cosners House Forum on LHC startup April 07 46 LHC operational cycle PHYSICS BEAM DUMP 9 14000 8 SQUEEZE PREPARE PHYSICS RAMP DOWN 6 START RAMP MB current 7 PHYSICS B [T] 5 PREINJECTION PLATEAU 4 3 INJECTION 2 T0 Ramp down 18 Mins 1 Pre-Injection Plateau 15 Mins Tinj 0 -3000 -2000 -1000 0 Time [s] Cosners House Forum on LHC startup April 07 1000 0 Injection2000 15 Mins 3000 Ramp Squeeze 28 Mins 5 Mins Prepare Physics Physics 10 Mins 10 - 20 Hrs 47 N 2 k b f L L F Eventrate / Cross TOT * kb f 4 n Stage I physics run Start as simple as possible Change 1 parameter (kb N *1 , 5) at a time All values for nominal emittance 7TeV 10m * in point 2 (luminosity looks fine) Parameters Beam levels Protons/beam ? 1013 (LEP beam currents) Stored energy/beam ? 10MJ (SPS fixed target beam) Rates in 1 and 5 Rates in 2 kb N * 1,5 (m) Ibeam proton Ebeam (MJ) Luminosity (cm-2s-1) Events/ crossing Luminosity (cm-2s-1) Events/ crossing 1 1010 18 1 1010 10-2 1027 << 1 1.8 1027 << 1 43 1010 18 4.3 1011 0.5 4.2 1028 << 1 7.7 1028 << 1 43 4 1010 18 1.7 1012 2 6.8 1029 << 1 1.2 1030 0.15 43 4 1010 2 1.7 1012 2 6.1 1030 0.76 1.2 1030 0.15 156 4 1010 2 6.2 1012 7 2.2 1031 0.76 4.4 1030 0.15 156 9 1010 2 1.4 1013 16 1.1 1032 3.9 2.2 1031 0.77 Cosners House Forum on LHC startup April 07 48 2 • • 1) 2) Physics inefficiencies What fraction of primary particles reach the various detector layers, versus pT? Charged particle component of MB events dominated by +-, with ~10% K+. First look at - Pixel system should provide 3 space points down to ~100MeV/c. Down to ~150MeV/c, particles can traverse SCT magnetic field, but reduced efficiency below 300MeV/c due to inelastic interactions in ID material. Cosners House Forum on LHC startup April 07 49 3 Physics inefficiencies What about the charged kaons? 1) 2) Much larger inefficiencies evident, due to increased material interactions and decays. For pT=300MeV/c, ~10% of K do not reach final Pixel layer. → 3) For pT=100MeV/c, K do not even make it into the Pixel system! In the SCT system, 15% of 500MeV/c K do not reach last SCT layer. → Below 300MeV/c no spacepoints in outer SCT. Cosners House Forum on LHC startup April 07 50