Impact Parameter Tag

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 1

Impact Parameter Tag

Use b-hadron long lifetime without reconstructing secondary vertex (to gain efficiency)

Consider compatability of “collection of tracks” (jet) with primary vertex

Use decay length significance of tracks:

vertex

Jet direction

Track direction

dca

Decay length

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 2

Steps to tag

1. Reconstruct jets, tracks, primary vertices2. Select best primary vertex for each jet3. Select tracks:4. Calculate PV probability for each track,

based on decay length significance distribution

5. Combine tracks in jet to form discriminant

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 3

Jet discriminant

Simplest approach: tag jet with a minimum number of tracks that have significance exceeding a certain value

Simple product of probabilities:D = -ln, = (Pi)…will yield larger values as #tracks increases

Default, independent of #tracks:D = -lnP,

in

1i

1n

0i

i

PΠΠ

!i/ΠlnΠPtracks

tracks

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 4

Track selection cuts

pT > 1.5 GeV DCA < 2cm (transverse and z) # SMT hits 2, total # hits 4 Track – Jet angle Track quality V0 rejection

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 5

Performance

1000 z->bb and 1000 z->ud samples ~5% fake rate at 50% efficiency for jets

within-2 < < 2, 3 tracks

Major improvement for vertical review mostly due to bug fix (sign error in decay length for some tracks)

Study track cuts, decay length error, 3D implementation for further improvement

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 6

Jet eta

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 7

Track pT cut

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 8

Track dca and nHits cuts

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 9

Track significance

L/dL (red: tracks with MCL > 0;black: tracks with MCL = 0)

|pT| < 1 Gev 1<|pT|<2 GeV

2<|pT|<5 GeV |pT| > 5 GeV

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 10

MC decL - Reco’d decL

|pT| < 1 Gev 1<|pT|<2 GeV

2<|pT|<5 GeV |pT| > 5 GeV

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 11

(MCL-RecoL)/dL

|pT| < 1 Gev 1<|pT|<2 GeV

2<|pT|<5 GeV |pT| > 5 GeV

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 12

Status

Sign bug fixed just before vertical review More performance improvement expected from

Better handle on error in jet / track angles Implementation of (pseudo-) 3D decay length Further study of track cuts (V0 rejection, 2?,

track direction) Other projects:

Jet acceptance Profiling of algorithm Higher statistics / certification sample studies

April 2001 Collaboration Meeting, 04-25-2001, Frank Filthaut & Bram Wijngaarden 13

<dump pics>