WZ Muon Selection For Winter Conferences

1
W/Z Muon Selection For Winter Conferences

• Tom Diehl
• FNAL
• _at_ Saclay 12/2001

Tight and Loose Muon selection for the WZ
group
2
Muon Quality

• Specification and Description of
• Tight
• Loose
• Astub
• What we have to do for these to be complete
• What I know
• What I dont know?
• W/Z -gt muons selection criteria
• Ideas I have or have heard that I liked

3
Muon Quality Tight

• Specification for Tight
• Identify all typical muons using muon-only
  information
• All muons available in one chunk All
  information about them available in that chunk or
  from pointers there.
• Uses the pieces of the muon detector that are
  also used in the L1 trigger
• Its used now in
• Reco for Muon-Corrected MET (Quality4)
• WZ_Reco for streaming from FARM
• La Macchina for selecting
• Muon_L3

4
Muon Quality Tight

• Definition of Tight MuonID
• Use MuonID chunk
• NS!hits(A-layer)gt0 (1 _at_ L1)
• NWhits(A-layer)gt1 (3-4 maximum)
• NS!hits(BC-layer)gt0 (1 _at_ L1)
• NWhits(BC-layer)gt2 (23-4 typical, 2-4 not
  unlikely holes )
• Chi2gt0 (fit converged)
• I believe it satisfies non-technical part of the
  specifications
• Tight definition is looser than in Run 1 - we
  can afford that
• The inner tracker will be a major contributor to
  muon ID, not only making the pT determination but
  also providing remaining background rejection

5
Specification for Loose

• Identify all muons that can be found with the
  muon detector
• not intended to be a muon-detector-unbiased
  sample
• keep in mind holes in the detector coverage and
  typical electronics failures
• Efficiency of Loose should be reliable with MC
  calculation
• Loose muons should provide a sample with which
  to measure the efficiency of each component of
  Tight
• Same technical specifications as for Tight
  i.e.- one chuck - all parameters available there
  or via pointers (we are stuck with tuples through
  winter conferences)

6
Muon Quality Loose

• Definition of Tight muon MuonID chunk
• 1a) NS!hits(A-layer)gt0 (1 _at_ L1)
• 1b) NWhits(A-layer)gt1 (3-4 maximum)
• 3) NS!hits(BC-layer)gt0 (1 _at_ L1)
• 4) NWhits(BC-layer)gt2 (23-4 typical, 2-4 not
  unlikely holes )
• 5) Chi2gt0 (fit converged)
• A Loose muon can fail one of the above criteria
  except 1a) and 1b) count only as one possible
  failure

Because A-layer scint. and drift chamber holes
are coincident
7
Drift Detectors
8
Scintillation Counters
9
Muon Quality Astubs

• Specification for Astubs
• primary Identify muons that dont penetrate the
  toroid magnets
• 1.1 lt pT(GeV/c) lt 3 or 4 in central
• 1.1 lt p (GeV/c) lt 3 or 4 in forward
• secondary Identify muons that miss all of BC

• Definition (hardly even preliminary)
• Astub is S!HitsWHits(A-layer)gt1
• For L3, its probably got overlap with Loose
  and Tight
• Warning Theres lots of these because of
  background

10
Problems with Quality Definitions
That I know of 15/Dec/2001

• In p10.13.00 all Tight are available in MuonID
  bank as nseg 3 or -3 (and some of the Loose
  are -3 as well) but Not ALL Loose are
  available in MuonID chunk.
• Loose must include BC tracks with a pT
  calculated from the vertex (nseg-2) as well as
  those that have a gtrk match (nseg2)
• Muon Duplication
• One real muon can occur in the MuonID bankchunk
  up to three (six) times if its in a dense
  region of central tracks, in a jet, perhaps.
• What else?

11
Muon Isolation Definition

• Most Physics Groups have agreed to use
  DR(mu-jets0.5)gt0.5 until someone demonstrates
  something better.
• We havent taken time to reoptimize this since
  Run I
• Note that central muon phi resolution is pretty
  bad (unless nseg is positive)
• Note that forward muon phi resolution is pretty
  bad (unless pixels are used in segment)

12
W/Z to Muons Selection

• Ws Tight Muon plus trackmatch
• Zs Probably Tight Tight Tight
  Loose and at least one trackmatch.
• Plus sensible kinematic requirements

13
Some Good Ideas

• Tight scintillator timing
• L1 trigger gate is 24 ns wide in central A-Layer.
  Efficiency would remain 99 if this was
  tightened to 18ns in L1 or L3 and probably could
  be tightened more Offline given some careful
  studies
• Similar comment for all the rest of the
  scintillator
• Count Decks instead of Hits in performing ID
  (Dave)
• PDTs MDTs have 3 or 4 decks
• This is obviously correct.
• Use the Inner Trackers !!!!!!!!!

14
Some More Good Ideas

• Locate the beam scraping source in the North
  Tunnel - a hot topic
• More Shielding in A-Layer - part of upgrade for
  Run 2b
• Muon unbiased muon
• I can imagine how we might do this and its
  something to think about six months down the road

15
Summary

• Ive discussed Tight and Loose muon
  definitions and their specifications
• W/Z group is applying a Tight muon definition
  and has a working definition for Loose
• We plan to select postshutdown events based on
  these definitions with the addition of Inner
  Tracker and sensible kinematic information.
• Good ideas that I have heard

16
Trailer

• MY Talk Ends Prior to Here

17
W/Z to Muons

• Single Muon Trigger Rates
• L4.7e30
• mu1cmsc_fz 46 (64/108) gt5.8 hz/1030
  (not including octants 5 6)
• mu1pix_fz 60 hz gt12.8
  hz/1030
• Readout Bandwidth seems to be 30 hz so single
  muons are getting prescaled heavily

Tighter gates
18
Efficiency Calculation

• Partly from Monte Carlo.
• Monte Carlo calculates acceptance kinematic
  selection
• Monte Carlo doesnt model the detector
  performance
• Partly from Data
• Use Data whenever possible
• See above in reverse for Data

19
Muon Changes I

• Reported by Frederic last time for p10.11.00
• muo_trackreco
• addition of quality word (quality
  4 corresponds to tight)
• muon_id
• addition of MuonQualityInfo
• int Nhit NWA10NWBC1000NSA10000NSBC
• method muoTrackIndex returns the local track
  index for nseg3 else -1
• method segIndex returns segment index for
  nseg12 else -1
• method TightMuoTrack returns T if tight muon
  (used by METcalmu)

20
Muon Changes II

• Migrating muon selection (e.g. tight loose)
  to Muon_ID
• Muon_ID
• nseg -3 criteria changed from NWAgt2 NWBCgt3
  0ltChi2lt1000 NWAgt0 NWBCgt1 -infltChi2ltinf to
  match nseg3 (note it takes 2 hits to make a
  segment)
• Failed to make an nseg -2 in muon_id because
  theres no BC-only tracks in muo_trackreco
  (Frederic is working on this)

21
Muon Changes II continued

• Muon_ID Track-Matching
• nseg 3 trackmatch in Central w/ PT(gtrak)gt3.0
  GeV/c (was 1.5 GeV/c) note its still 1.5 GeV/c
  in Forward note its 3.0 GeV for BC-only
  (nseg2) note its 1.5 GeV/c for Astubs
• Criteria dZ(A-L.cm) dPhi dEta
• SMT (100/pt)10 0.5 0.4
• Global same same same
• CFT none same none
• (all now RCP controlled) (was 4)
• Still, because in Muon_ID we match on A BC
  segments, its possible to have an A-stub or
  BC-seg with a trackmatch in common with an ABC
  muon?
• Track-matching is done with a CH loop
  (Pt-ordered) priority-ordered by muo_trks,
  BC-segs, Astub-segs

22
Muon L3 Filter

• What criteria are we going to have available to
  identify muons in L3?

• We separated muon object from object
  relational criteria and concentrate on the muon
  object to begin with.
• In step with the offline selection.
• Providing muons well-matched in selection
  criteria
• Providing means to measure efficiency
• Attempting to anticipate the requirements of
  triggers

23
Muon Selection Criteria
First Try

• Muon_L3(key,Nmu,eta,Region, Quality,PtLocal,TrackM
  atch, PtTrackmatch,CalMatch)
• key is a marker for relational tools
• Nmu is the required count
• eta cut (e.g. eta(muon)lteta)
• Region (e.g. forward/central)
• Quality (Tight ,Loose, Astub) based on
  muon criteria more coming up
• PtLocal (Pt of the muon local track)
• Trackmatch criteria more coming up
• PtTrackmatch (Pt of inner tracker)
• CalMatch criteria not much yet

24
Track Match Criteria

• Matching with Central Tracks is important for
  muons, especially since we anticipate (and see)
  relatively pure muon triggers.
• Capability in hand includes global trackmatch via
  best spatial criteria.
• Paul is expanding the tracking detector options
  and
• is considering selecting the track with a
  Pt-ordered spatial criteria.