Connecting the dots… - The DZero Experimentjabeen/index_talks_files/kansas_talk.pdfConnecting the...

Connecting the dots…(Tracking at D0)

Shabnam JabeenKansas University

5/6/03 Shabnam Jabeen (Kansas) 1

IntroductionTracking part of D0 DetectorTracking AlgorithmsUse of Tracking AlgorithmsPerformance

bdfrtvv

The Question is..

Where did all this come from?What is all this made of?

What was universe like in the beginning?

How many fundamental particlesand forces there are?

why particles have charge?

why particles have mass?Why ------ -- -- -- ?

How …….?

…. .. .. .. .. ..


STANDRAD MODELOne of the most beautiful, rich in predictionsand experimentally verified models given ever

No Higgs yet !

Remarkable agreement between theory and experiment up to energies of ~ 100 GeV

But ………


We are still far from

Theory of EVERYTHING

Large number of theories out there but ….


“Where would a brave knight be without his noble steed ?”*

A theory is just a beautiful thought until it is proven experimentally

* It does not mean that a theorist is an Ogar and an experimentalist is a donkey , or vice versa


Experimental tools• Accelerators, detectors, computers

• Theoretical and simulation tools

• Analysis techniques


High Energy Research is mainly carried out with large high High Energy Research is mainly carried out with large high energy energy colliderscolliders and complex particle detectors in few and complex particle detectors in few international laboratoriesinternational laboratories



Tevatron – Highest energy accelerator at present

Tevatron

Higher energy collisions: 980 GeV

Number of p and pbar bunches: 36 (396 ns beam crossing time)

Luminosity: Run 2 goal 2-4 x 1032 cm-2 s-1


980 GeV 980 GeV

P Pbar


Modern Detector


DZERO Detector

Coarse Hadronic(Tail catcher)

Fine Hadronic

EM

DØ

Muon system

Magnetized iron

Tracker

5000 tons of iron, uraniumLiquid argon, coils, cables, wires, …


Calorimeters Tracker

Muon System

BeamlineShielding

Electronics

protons antiprotons

20 m


DØ Tracking SystemSilicon Tracker793,000 channels

Fiber Tracker77,000 fibers

Solenoid2T superconducting

Central Preshower6000 channels

Forward Preshower16,000 channels


Silicon Microstrip Tracker

SMT half cylinder

Beryllium bulkhead

Charged particle momentum and secondary vertices

Provide tracking out to |η| ~ 3 Radiation hard > 1 MradMaximum silicon

temperature < 15 C


Silicon Microstrip Tracker

6 Barrels12 Central F-disks4 Forward H-disks3m2 silicon, 793k channels SVX chip readout

H diskbarrelF disk1.2 m

p

⎯p

10.5 cm

26 cm


Barrels and Disks


SMT Readout System

SVX IIe Chip• 128 channel, • Radiation hard• Analog pipeline (32 cells)to store signals

While L1 trigger is formedp-side pulse-height (ADC)ADC counts

1 mip ~ 4fC ~ 25 ADC counts. Noise < 2 ADC countsS/N > 10 as expected


Scintillating Fiber TrackerEight cylinders covered withscintillating fiberare read out with

a novel light detector (VLPCs).

Two main functions1.Track reconstruction and momentum

measurement with silicon system 2.Fast Level 1 Triggering

|η| ~1.7Hit resolution about 100µm


Fiber TrackerBarrels

8 carbon fiber barrels20<r<50cm

side view

end view

axialstereo

axialstereo

Scint Fibers830µm diameter, 2.6m length

Fiber Ribbons256 fibers8 axial doublets8 stereo doublets (2o pitch)


D3-G

4

B1

5A 5B

E3

R99.000

A2-G

5

A2

G5

C3

E4

C3-E

4

A1

H5

A1-H

5

B2

F4

123

4 5

B2-F

4

274

296

232

135

177

B3

H6

B3-H

6

D4-F

5

D4

F5

B5

B5-H

11

F9

G10

H11

F9-G

10

D7

E8

D7-E

8

C6

H10

B1-E

3

C2

F3

H4

D2

C2-H

4

D2-F

3

D3

G4

C6-H

10

G9

A4

A4-G

9

E7

G3

H3

G3-H

3

E2-F

2

E2

F2

F8

E7-F

8

D6

H9

D6-H

9

C5

G8

C5-G

8

E6

F7

E6-F

7

B4

H8

B4-H

8

G7

D5

D5-G

7

A3

F6

A3-F

6

C4

H7

C4-H

7

E5

G6

E5-G

6

C1-D

1

C1

D1

E1

F1

G1

H1

H2

G2

E1-G

1

F1-H

2

G1-H

1

C.P.

S.

C.P.

S.

H12

H13

H14

H15

H16

H17

G11

G12

G13

G14

G15

F10

F11

F12

F13

F14

E9

E10

E11

E12

D8

D9

D10

D11

C7

C8

C9

B6

B7

B8

A5

A6

H18

H19

H20

H21

H22

G16

G17

G18

G19

G20

G21G22

F15

F16

F17

F18

F19

F20

E13

E14

E15

E16

E17 E18

D12

D13

D14

D15

D16

C10

C11

C12

C13

C13

B9

B10

B11B12

A7

A8

A9

A10

B13

C14

C15

D17

D18

E19

E20

F21

F22

F23

G23

G24

G25

H23

H24 H25

H26

H27

H28

B6-C

8-F1

0-H1

2E9

-G11

Clear waveguides8-11m clear waveguide tophotodetector

Fiber Tracker

Scintillating fibers (or strips)

Light detector (visible light photon counters VLPC)– Necessitate

CryogenicsReadout electronicsTrigger

~9K

Scintillating fiber (CFT) orwavelength shifting fiber (PS)

Opticalconnector

Clear fiber waveguideVLPC

cassette

AFE Boards

Cryostat

Mirror


Central and Forward Preshowers


Central mounted on solenoid (|η| < 1.2) CPS: 7,680

Forward on calorimeter endcaps(1.4 < |η| < 2.5)

FPS: 14,000 channels

Solenoid

Central PS

CFT

VLPCs for DØThe DØ detector uses VLPC readout for the

following subsystems:The scintillating fiber trackerThe central and forward preshower

VLPC: Visible Light Photon CounterSolid state photon detectorsDetects single photonsCan work in a high rate environment Quantum efficiency ~80%High gain ~40 000 electrons per converted

photonLow gain dispersionoperate at :

8 K < T < 10 K, 6 V < Vbias< 7.5 V+ -

IntrinsicRegion

GainRegion

DriftRegion

SpacerRegion

Photon

e h

Substrate


VLPC Readout Electronics

AFE boards provideBias voltage + Temperature control for VLPCFast discriminator digital trigger information for L1CTTAnalog pulse information (SVXII chip)512 channels per board

8 MCM each with 4 SVX + 1 SIFT chipSEQ + VRB readout as SMT

0 pe

2 pe

3 pe4 pe

1 pe1 pe ~ 7 fC ~ 15 ADC counts

Expected signal for MIP ~ 8 pe in CFTExcellent signal/noise performance

ADC counts


Solenoid

2 Tesla, ±4820 Amps 5.6 MJoules stored

energy4.5K cooling2.7 m lengthField uniformity better than

0.5%


Detector

Amplifier

Digitizer

selection

storage

computers

Particle

signal

Trash

010010


Tracking

Ω→ΛKFind association between hits (detector measurements)Construct tracks(particle trajectories )using hits


Tracking Algorithms Global tracking or Road approach: Uses specific paths (roads) during track findingAuthor: GTR group

HTF (Histogramming Track Finder): Divides D0 detector into slices in Can use either CFT, SMT hits or their combinations to construct tracksAuthor: S. KhanovElastic Reco(Elastic Template Algorithm): Can use existing tracks as initial seeds, existing vertices or run in stand alone modeAuthor: A. Haas

AA (Alternative Algorithm): Starts from 3 hits in different superlayers (SMT or CFT).Track candidates are extended towards CFTAuthor: G. Borrisov

( , )ρ ϕ

Variety of Tracking Algorithms allow cross-check the tracking performance

Select the best one


Tracking Algorithms Object Oriented approach (C++)

Variety of Tracking Algorithms allows cross-check the tracking performance

May be run simultaneously/separately to achieve best performance for different physics tasks

All of them use one final track refitting step

One common user interface


Global Tracking Surfaces: Build a model of the tracking detectors using abstract surfaces. Cylinders for fiber fiber tracker, x, y and z planes for silicon detectors

Paths: An ordered list of the surfaces that a particle coming from a ppbar collision would cross. First few surfaces are used to build a seed track.

Propagators: Are used to extrapolate the seed tracks between the remaining surfaces. It solves the equation of motion for a track, including the effects of magnetic field, multiple scattering and energy lost in the material

Fitters: Once the track reaches a new surface, fitter attempts to add a new cluster to the track

Filters: After moving through all surfaces in a particular path, a number of filters are applied to clean the list of candidate tracks


Global Tracking

The final output of each GTR path is a list of tracks

Tracks themselves store a list of surfaces used and the track parameters at each surface


So we have connected the dots!

7.5M p-pbar crossings every second2 TB/sec data volume

Hadron collisions:Very messy Hundreds of objects after collision

Need to simplify the measurement


DØ Trigger Overview

was 300 kHz in Run I


-

Hardware

Hardware/Software

PC’s & C++ algorithms

was 3 Hz in Run I

L1: Synchronous

L2: Asynchronous

L3: PC Farm

Data disk

10 kHz

1 kHz

30-50 Hz

7.5 MHz

increase in readout precision& object quality

-

4.2 µs

100 µs

50 ms

250 kBytes/event

DØ L1 & L2 Triggers

Central tracking system plays an important role in L1, L2 trigger

CAL

c/f PS

CFT

SMT

MU

FPD

L1Cal

L1PS

L1CTT

L1Mu

L1FPD

L2Cal

L2PS

L2CTT

L2STT

L2Mu

Global L2Framework

Detector

Lumi

Level 1 Level 27 Mhz 5 khz 1 khz

Level 3


Use of Tracking Algorithms Example: Hit finding efficiency for CFT for raw data

Detector

Amplifier

Digitizer

selection

storage

computers

Particle

signal

Trash

010010

Take Raw data from storageReconstruct the Tracks using GTRUse these tracks and associated information to find CFT Efficiency


Use of Tracking Algorithms

Track selection

• Number of CFT hits =15• Pt > 1 GeV• No more than 1 hit within 11 σ (Tracking)

• Window = 3 σ (Resolution)

σ(Tracking) ~ 100 µσ (Resolution) ~ 0.013 cm

xx

x x

x

xx

RoadWidth(11σ )

Window

Skipped layer

cluster

Example: Hit finding efficiency for CFT for raw data


Use of Tracking Algorithms Hit finding efficiency for CFT using raw data


Detector Performance: Tracking

p-side pulse-height (ADC)

Silicon cluster charge

1 mip ~25 ADC #Noise < 2 ADC #

track pseudorapidity

fiber light yield

Scintillating Fiber Tracker (CFT)first time in a collider detector

• performs as expected • ε > 98% (including dead channels)

ADC counts

Silicon (SMT)• good S/N• ε>97% (incl. dead chnnels)


Detector Performance: Tracking

Almost on target with• no CFT alignment• 1st pass SMT alignment


Impact Parameter

track

x

y

DØ Run 2 Preliminary

width =36µbeam ~30µ

DØresolution:

~20µ

Proper decay length (cm)-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3

mµE

vent

s/50

1

10

102

103

Proper decay length (cm)-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3

mµE

vent

s/50

1

10

102

103

Data

Background from sidebands

ψBackground + prompt J/

B lifetime signal

Signal+background

D0 Run II Preliminary

+ X)ψ J/→Average B lifetime (B Proper B decay length (B J/ψ+X)

B life time signal

Prompt J/ψ

Decay length (cm)

Data

)2

mass (GeV/c1 1.5 2 2.5 3 3.5 4 4.5 5

2N

eve

nts

/ 10

0 M

eV/c

0

10

20

30

40

50

60

Run 2 Preliminary∅D

-e+ e→ ψJ/J/ψ→e+e-

Muon System

Calorimeter

J/ψ and ψ’

E/P0 0.5 1 1.5 2 2.5 3

Eve

nts

/ 0.0

5

0

50

100

150

200

250Chi2 / ndf = 0.5317 / 4

156.4 ±p0 = 105.6

490.4 ±p1 = -411.3

459.2 ±p2 = 530.6

141.7 ±p3 = -192.2

26.31 ±N = 147.6

0.01741 ±Mean = 1.031

0.02659 ±Sigma = 0.165

DØ Run 2 Preliminary

candidate eventsν e→W

>20 GeVmissT>20 GeV, Eel

TP

E/P Chi2 / ndf = 0.5317 / 4

156.4 ±p0 = 105.6

490.4 ±p1 = -411.3

459.2 ±p2 = 530.6

141.7 ±p3 = -192.2

26.31 ±N = 147.6

0.01741 ±Mean = 1.031

0.02659 ±Sigma = 0.165

W→eν

E(cluster)/p(track)

2GeV/c3.5 3.55 3.6 3.65 3.7 3.75 3.8 3.85

0

10

20

30

40

50

mass -µ +µ mass -µ +µ

Combine tracking with other systems

Detector Performance


Z→µ+µ-

~170 evts

What are we going to find?

I don’t know?

But whatever it is, we are not going to miss it!


Track Parameters


Central Track Trigger


DFEA

DFES

DFEF

Silicon Track TriggerL1CTT finds tracks in CFTFiber Road Card

– Receives tracks from L1CTT and trigger info

– transmits trigger road infoto STC+TFC

Silicon Trigger Card– Select SMT clusters in roads– Receives raw data from SMT– Finds clusters in axial and

stereo strips– Associates tracks with axial clusters

Track Fit Card– Take FRC roads and STC clusters and fit

track trajectory– Output Track list to L2

roaddata

SMTdata

SiliconTriggerCard

SiliconTriggerCard

SiliconTriggerCard

SiliconTriggerCard

SiliconTriggerCard

FiberRoadCard

SiliconTriggerCard

TrackFit

CardL2CTT

±1 mm


Liquid Argon CalorimeterD0 LIQUID ARGON CALORIMETER

1m

CENTRAL CALORIMETER

END CALORIMETER

Outer Hadronic(Coarse)

Middle Hadronic(Fine & Coarse)

Inner Hadronic(Fine & Coarse)

Electromagnetic

Coarse Hadronic

Fine Hadronic

Electromagnetic

Drift time 430 ns

p

p

Z

y

x

θ

ϕ


Liquid argon sampling– Stable, uniform response, rad. hard– LAr purity important (< 0.7 ppm O2 equivalent)

Uranium absorber (Cu/Fe for coarse hadronic)– dense absorber hence can be compact– Nearly compensated EM and hadronic response– Linear response

Hermetic with full coverage– |η| < 4.2 (θ ≈ 2o)

λ int > 7.2 (total)

Muon Detector


Bottom B/C Scint

Central Trigger Scint (A-φ )PDTs

ForwardTrigger Scint(Pixels) A,B,C

Forward Tracker (MDTs)A,B,C

Shielding

• Muon rapidity coverage to eta of 2.0

• Shielding reduces backgrounds by 50-100x


“passive” sensor

“active” sensor

SVX2e readout chips

HDI (flex circuitreadout)

Wire bonds

Two b-jets fromHiggs decay

Missing ET

Electron Track

EM cluster

CalorimeterTowers

p → ←⎯p

⎯pp → WH →⎯bb

→ eν

Hits in Silicon Tracker(for b-tagging)


Pipeline Control Logic

AnalogPipeline:

128 channel

32 cells128

Inte

grat

ors

AD

C C

ompa

rato

rs

Spar

sific

atio

nFI

FO

I/O

ADC rampand

counter

Analog section Digital section

To S

ilico

n D

etec

tor

To R

eado

ut S

yste

m


Tracking

Complications: low momentum tracks, scattering, noise,high track density in jets, etc., etc., etc.

Need: Balance between CPU, reconstruction time and efficiency Vs. scattering, noise, pT threshold

Different Tracking Algorithms


Connecting the dots… - The DZero Experimentjabeen/index_talks_files/kansas_talk.pdfConnecting the...

Documents

Transcript of Connecting the dots… - The DZero Experimentjabeen/index_talks_files/kansas_talk.pdfConnecting the...