The ALICE Transition Radiation Detector Design and Performance

detector principle and overview results from testbeam measurements

dE/dx transition radiation electron/pion separation position and angular resolution

performance of electronics status of the project

Johannes P. Wessels, Universität Münster for the ALICE TRD Collaboration

IEEE Nuclear Science Symposium, Rome, Oct. 16-22, 2004

ALICE SetupALICE setup

ITSTPC

TRD

TOFPHOS

HMPID

MUON ARMPMD

FMD

Transition Radiation Detector (TRD)

Purpose:electron ID in central barrel p>1 GeV/cfast trigger for high pt particles

Parameters:540 modules -> ~760m2

length: 7manticipated X/X0 ~ 15 %

28 m3 Xe/CO2 (85:15)1.2 million channels17 million pixels15 TB/s on-detector bandwidthweight ~ 21 ttotal power: ~ 60kW

Principle of Operation

two purposes: PID & momentum measurement

7 mm

31 mm

48 mm

Radiator

polypropylene fibers (~17 m) CF-backed ROHACELL foam irregular sandwich radiator parameterized for simulations

Full Size Radiator

size: 1200x1600 mm2 deformation at center 1 mm @ 1 mbar

Mounting of Electronics

Radiator Comparison

method: likelihood on total charge averaged over 4 detectors

extrapolated to six layers

pion rejection of 100 achieved over large momentum range

little dependence on actual radiator producer

Improvement of -rejection

bi-dimensional likelihood analysis improves pion rejection

probability of finding largest cluster in a given time bin

-Rejection Using Neural Network

feed-forward neural network w. 15 input neurons

2 hidden layers

factor 3-7 improvement over LQX method

pion rejection of 200-500

-Rejection vs. incident angle

slight deterioration of pion rejection at small angles (0o-2o)

not frequent in ALICE space charge effects

diminish signal not included in

simulations low gas gain preferable

Resolution vs. Incident Angle

quantitative understanding of all resolution effects

significant improvement in position resolution with tail merging and tail cancellation

position resolution better than 300 m

angular resolution better than 0.8

o

Resolution vs. Signal-to-Noise

resolution better for pions at given S/N ratio

average signal larger for electrons

comparable resolution for electrons and pions

angular resolution smaller for electrons with radiator -> L-shell fluorescence

TRD electronics chain

PASA TRAP - digital chip

40mm

Preamp Shaper (PASA)

18 4th order preamplifier/shapers with differential outputs (21) 12 mV/fC, 13 mW/channel

digital test structure for chip verification size of chip: 3030 µm x 7280 µm full production received; thinned to 300 µm

PASA – test results

gain: 12.2mV/fCdynamic range: 0.15fC..165fCshaping time: 40nsFWHM: 120nsdifferential output: -1..+1Vnoise at 25pF: 702enoise slope: 21e/pFintegral non-linearity: <0.16%power consumption: 13 mW/channel

crosstalk as function of inter-pad capacitance

ADC Performance

Muthers, Tielert, KaiserslauternMuthers, Tielert, Kaiserslautern

0.18 m CMOS10 bit, 10 Ms/s0.1 mm2, 9.5 mWENOB: 9.5 @ 1 MHzDNL: -0.4;0.6INL: -0.8;0.7

Filter

Non-linearity

Pedestal

Gain

Tail

Crosstalk

TRD Trigger Timing

Drift time

59901996998 2994 3992 4990 t [ns]

relevant pipeline ADC output

Calculate fit

PASA ADCTracklet

Preprocessor

TPPTRD

event buffer

Calculate Tracklets

TrackletPreprocessor

TPP

TrackletProcessor

TP

event buffer

TrackletMerger

TM

Data ship

Global Tracking

GTU

TRD TRD

N. Herrmann, V. Lindenstruth, B. Vulpescu

TRD Stack Preparation

test of 6 chambers at CERN this week

e/ - beam up to 10 GeV/c

Cosmic Ray Track

readout with MCMs Ar/CO2 (85/15) Vanode = 1400 V vd = 2.6 cm/s

Summary

pion rejection and tracking capability fulfill specs quantitative understanding of

dE/dx, position and angular resolutionTR production & absorption

promising results of PASA and digital ASIC evaluation

good trigger capability for high pt charged particles

starting series production now aim to be ready for first events in 2007 physics performance report http://alice.web.cern.ch/ALICE/ppr

ALICE TRD Collaboration

C. Adler, A. Andronic, V. Angelov, H. Appelshäuser, C. Baumann, T. Blank, C. Blume, P. Braun-Munzinger, D. Bucher, O. Busch, V. Catanescu, V. Chepurnov, S. Chernenko, M. Ciobanu, H. Daues, D. Emschermann, O. Fateev, S. Freuen, P. Foka, C. Garabatos, H. Gemmeke, R. Glasow, H. Gottschlag, T. Gunji, M. Gutfleisch, H. Hamagaki, N. Heine, N. Herrmann, M. Inuzuka, E. Kislov, V. Lindenstruth, C. Lippmann, W. Ludolphs, T. Mahmoud, A. Marin, J. Mercado, D. Miskowiec, Y. Panebratsev, V. Petracek, M. Petrovici, C. Reichling, K. Reygers, A. Sandoval, R. Santo, R. Schicker, R. Schneider, S. Sedykh, R.S. Simon, L. Smykov, J. Stachel, H. Stelzer, H. Tilsner, G. Tsiledakis, I. Rusanov, W. Verhoeven, B. Vulpescu, J.W., B. Windelband, C. Xu, V. Yurevich, Y. Zanevsky, O. Zaudtke

Physikalisches Institut, Universität Heidelberg, Germany; GSI, Darmstadt, Germany; Kirchhoff Institut, Universität Heidelberg, Germany; FZ Karlsruhe, Germany; Universität Frankfurt, Germany; Universität Münster, Germany; NIPNE, Bucharest, Romania; JINR, Dubna, Russia; University of Tokyo, Japan

Transition Radiation

Poisson distribution for measured TR photons

some loss of TR clusters in analysis

2 GeV/c

Parameterization as regular foil stack

270 interfaces, 10 m thick, 80 m spacing

no momentum dependence in simulation