Transcript B tagging at L2 - INFN - Istituto Nazionale di Fisica Nucleare

The Silicon Track Trigger
(STT) at DØ
Beauty 2005 in Assisi, June 2005
Tag
beauty
Sascha Caron
for the DØ collaboration
fast …
The Silicon Track Trigger at DØ
OUTLINE
- D0 detector
- D0 trigger
- What is the STT?
- dataflow
- clustering and tracking
- B identification
- Performance
- Summary
Sascha Caron
page 1
The Silicon Track Trigger at DØ
D0 in Run II
The Silicon Track
Trigger is based
on information
of the :
Silicon Microstrip
Tracker
Central Fiber
Tracker
Sascha Caron
page 2
The Silicon Track Trigger at DØ
D0 in Run II
The Silicon Track
Trigger is based
on information
of the :
Silicon Microstrip
Tracker
o 50 μm pitch strips (parallel to beam)
o 800000 readout channels
o Only axial strips used for STT tracks
Central Fiber
Tracker
o Scintillating fibers
o Coverage: |Rapidity| < 1.7
Sascha Caron
page 3
The Silicon Track Trigger at DØ
Events 10
Aim: High ET
per
Z->bb, HZ->bbvv,
second
H->bb, etc.
1
Hopefully B physics
Find b-events
early to keep
high efficiency
at an acceptable
rate
0.1
0.01
QCD ET>30 GeV
dijet production
b-jets ET>30 GeV
Z-> b bbar
0.001
Events for
L=100 E30 cm-2/s
= about 8 pb-1/day
0.0001
Sascha Caron
Higgs->b bbar
ZH-> bbvv, bH->bbb etc.
page 4
The Silicon Track Trigger at DØ
Principal Idea
B decay products
Interaction point
is mean beam spot
Impact parameter
(2d in x-y plane)
o Silicon Improved Tracks with
2d impact parameter
o Select events with large impact parameter tracks
Sascha Caron
page 5
The Silicon Track Trigger at DØ
Trigger System
p¯
p bunch
crossing
frequency
2.5MHz
L1 Trigger
(1500)
5000 Hz
decision time
about 4 μs
o Hardware based
o CTT tracks,
calorimeter towers,
muons
Sascha Caron
L2 Trigger
(700)
1000Hz
L3 Trigger
50 Hz
decision time
about 200 μs
decision time
about 50 ms
o Hardware/Software
o simple jets, electrons,
muons, taus
o Silicon Microvertex
improved tracks (STT)
o global processor to
combine information
(e.g. STT tracks for Bid)
o Software based
o partial event
reconstruction
page 6
The Silicon Track Trigger at DØ
Data flow chart L2 tracks
Central
Fiber
Tracker
L1
CTT
L2
CTT
other
L2 preprocessors
L2
Global
L2
STT
Silicon
Micro
Detector
Detector
Level 1
Sascha Caron
Level 2
page 7
The Silicon Track Trigger at DØ
How is the
tracking improved?
o Tracks found at L1 with the Central Fiber
Tracker are used to define roads into the
Silicon
o Silicon hits are clustered
o Track is re-fit within the road
Sascha Caron
page 8
The Silicon Track Trigger at DØ
Silicon Track Trigger cards
Road data to
Fiber Road Card
Fitting done
in Track Fit Card
L2CTT
Done in parallel for each
of the 12 sectors
SMT data to
9 Silicon Trigger
Cards
Sascha Caron
2 sectors in 1 crate
-> 6 Crates with
1 Fiber Road Card
9 Silicon Trigger Cards
2 Track Fit Cards
page 9
The Silicon Track Trigger at DØ
Silicon Track Trigger cards
Road data to
Fiber Road Card
- Receives and distributes L1
tracks
- Communicates with Trigger
framework
Fitting done
in Track Fit Card
- receives road and axial clusters
- convert to physics coordinates via LUT
- perform track fit φ(r)= b/r+kr+φ0
- beam spot correction
- output tracks to L2CTT
Sascha Caron
SMT data to
Silicon Trigger
Cards
- Perform clustering and clusterroad matching
- Clusters SMT hits + pedestal
correction
- Axial clusters are matched to
the roads
page 9
The Silicon Track Trigger at DØ
Performance studies STT tracks
Correlation to RECO
Resolution in μm
Impact Parameter resolution
pT(GeV)
Purity
Including beam spot size 35 μm
and 15 μm SMT resolution
Sascha Caron
page 10
The Silicon Track Trigger at DØ
B identification
Store 2d signal (b events) to
background (light quarks)
ratio into a lookup table
Ri
Signal to
Background
ratio
 Define “good tracks”
 Loop over tracks i and
derive the product
R= ∏Ri
or use max(Ri)
 Likelihood as a
discriminator: L= R/(R+1)
Sascha Caron
- Fast, simple and efficient technique
- Time needed is 2 operations per track and goes
linear with the number of tracks -> about 1-5 μs
- Degrades bad tracks without cutting them
page 11
The Silicon Track Trigger at DØ
Performance studies
L2 B-id likelihood
efficiency Z->bb
Red :
Simple Cut
Blue/Green:
B-id
Very well separation of signal and
background events
Sascha Caron
Large background reduction
with L2 b-ID algorithms
page 12
The Silicon Track Trigger at DØ
First studies
using data
DØ
preliminary
Using MC pdfs
Jet data
Offline B-tagged
data
DØ
preliminary
Likelihood shifted to
higher values due
to e.g. min bias effects
Max IP Significance
Likelihood
L2 data well correlated with offline data
-> proof of b-id principle -> Much more studies on their way
Sascha Caron
page 13
The Silicon Track Trigger at DØ
Summary
o
o
o
o
o
The STT is operating routinely and smooth
High Impact Parameter resolution
Data is understood
Very fast L2 b-id is (almost) implemented
We expect a high potential for
Z->bb and Higgs physics now and at L=200E30cm-2/s
Outlook
o Implementation of new Layer 0 (r=22mm) into STT
-> further increase in precision and stability
Sascha Caron
page 14