1

Forward Tracking Simulation

Norman A. Graf

ALCPG Cornell

July 15, 2003

2

OverviewDevelop track-finding strategies for the forward

disk regions in SD.Detectors and support material defined.Ideal MC hits written out.Detector digitization post facto• Study various readout schemes (pixel, stereo angle)• Hit merging and ghosting important.

Include beam backgrounds• Largest source of hits!

3

Tracking Strategy Find tracks in the 5-layer CCD pixel VXD,

extrapolate forward to pick up hits in the disks.

Conformal mapping of 3D hits simplifies pattern recognition.

Find tracks from outer radius in, fit, then extrapolate fitted track outwards.

Use Kalman filter to reconcile track hit prediction with measured strips.

4

Tracking in VXDPattern recognition for well-measured,

separated 3D points is not a problem.Five layers provide sufficient redundancy.Test pattern recognition in simplified events

events: 1, 10, 100, 200, 500 /event3, 5• 4 < < 176• 1GeV < E < 10GeV

5

Pattern RecognitionConformal-mapping technique applied to 3D

hits in VXD and forward disks.• Hits smeared by expected resolutions:• 5 in r and z for CCD

• 7 in r and r for FWD

• No hit merging!• No ghosts!

Treat as combined system:• Find VXD+FWD tracks in forward region.

6

1000 Events 500/Event

-3 -2 -1 0 10

100

200

300

400

500

600

700

800

Found-ShouldHaveFound

Missed ~250/500,000 (99.95%)

7100 200 300 400

1

10

100

1000 mean : 35.172 rms : 41.218

Tracking Time (miiliseconds)

Track-Finding Time/Event

500 /event700MHz

PIII

8

Hit-Merging in VXDCurrently record exact position of MC track’s

intersection with sensitive volume in simulations.

Smear with expected measurement resolution• Default is 5 microns.

Hits are currently distinct, even when they are within a pixel (20 microns!).

Real hits populate ~2×1 set of pixels.SLD experience being studied.

9

Adding BackgroundsBackgrounds arising from pairs hitting the

beampipe have been generated and passed through the full simulation packages.

One can overlay such events from an arbitrary number of beam crossings onto signal events.

10

Single Beam Crossing

11

Single Beam Crossing

12

Hit Densities vs. Radius

SD

T. Maruyama

13

0 50 100 150 2000

100

200

300

400

500

600

700

800

900

1,000

1,100

1,200

1,300

1,400

1,500

1,600

1,700

1,800

1,900

2,000

2,100 mean : 37.556 rms : 21.329

VXDHits - Layer1 number of hits

0 20 40 60 80 100 120 140 1600

100

200

300

400

500

600

700

800

900

1,000

1,100

1,200

1,300

1,400

1,500

1,600

1,700 mean : 62.165 rms : 20.584

VXDHits - Layer0 number of hits

0 20 40 60 80 100 120 1400

200

400

600

800

1,000

1,200

1,400

1,600

1,800

2,000

2,200

2,400

2,600

2,800

3,000

3,200

3,400

3,600

3,800

4,000 mean : 13.867 rms : 12.765

VXDHits - Layer2 number of hits

0 20 40 60 80 100 1200

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

5,500

6,000

6,500

7,000

7,500

8,000 mean : 6.1518 rms : 7.8351

VXDHits - Layer3 number of hits

14

VXD TrackingSystematic investigation of pattern

recognition in presence of full backgrounds not yet completed, but no problems experienced yet.

Effects of hit-merging to be studied.• not expected to be a major factor.

Efficiencies and resolutions as a function of and momentum being compiled.

15

Forward Tracking1. Extrapolate found tracks to hit strips.

2. Associate strip hits (either double-sided or back-to-back single-sided) in wedges of z-disks to form 3D space points.• Need systematic study of occupancies for various

designs.• Can we survive the ghosts? Grow as ~n2!

3. Use pixel hits if available. Detailed hit merging and ghosting needed!

16

Forward Disk DetectorsMany open issues:• Tiling of disks with wafers:• Phi segmentation?

• Radial segmentation?

• Mix of Si pixel and -strip detectors?• If pixel, APD or CCD?• If -strip, double-sided or back-to-back?• Strip orientations within wedges.• Shallow- or large-angle stereo?

17

Tiling Forward Disks

18

Strip Orientations

Shallow Angle Stereo Large Angle Stereo

19

Next StepsCurrently developing flexible tools to study

effects of disk tiling and strip orientations.Will start characterizing various detector

layouts in terms of efficiencies and resolutions.

Continue with impact on physics channels of interest for forward regions.

20

SummaryStrategies being implemented to handle

pattern recognition in the forward regions.Detector digitization infrastructure (hit

merging and ghosts) is almost in place.Recognize that detector design requires

reconstruction input.

Aim for flexible framework to allow iteration.