The Time Projection Chamber for the ALICE Experiment Luciano MUSA CERN, Geneva, Switzerland -

Transcription

1 The Time Projection Chamber for the ALICE Experiment Luciano MUSA CERN, Geneva, Switzerland - luciano.musa@cern.ch The Alice Time Projection Chamber a Technological Challange in the LHC Heavy Ion Physics Luciano Musa (CERN) The ALICE Time Projection Chamber Outline General condition at LHC for heavy ion collisions The TPC in the ALICE Detector Luciano Musa CERN The ALICE TPC Collaboration Challenges at high particle density TPC Main Components Field Cage Readout Chambers Electronics Commissioning the full TPC Summary and Outlook Bergen CERN Darmstadt TU GSI Darmstadt Heidelberg PI Lund Bratislava Copenhagen Frankfurt Heidelberg KIP Krakow General Conditions at LHC for Heavy-Ion Collisions The ALICE Detector ALICE, a general purpose Experiment measures hadrons, leptons and photons at mid-rapidity Pb Pb: 5.5 TeV CM-energy (NN) pp, pa, A-A «The biggest step in energy of the history of heavy-ion collisions» G. Rolland HMPID PID high p t TOF PID p t = -3 GeV / c Other detectors not shown FMD, V, T, ZDC Luminosity (max) Pb + Pb:. 7 [cm - s - ] 8 khz interaction rate event (central) rate Hz Rapidity density predictions dn ch / dy = 6 (model dependent) What can we learn from RHIC? The first LHC event will give an answer p + p: 5. 3 [cm - s - ] khz interaction rate event rate > khz ALICE Detector designed for dn ch / dy = 8 dn/dy 35 at = educated extrapolation (saturation model, Eskola et al.) TRD Electron ID PMD multiplicity ITS Low p t tracking Vertexing TPC tracking, de / dx PHOS, TPC tasks track finding momentum measurement particle identification. GeV/c < p t < GeV/c <.9 Requirements tracking efficiency: > 9% momentum resolution: <.5 % MUON - SPECTROMETER de/dx resolution: < % Detection and Identification of pairs two track resolution: < 5MeV/c rate capability: Hz central Pb-Pb ( KHz p-p) 3 4 TPC Overview Challenges at high particle multiplicities High Voltage electrode ( kv) E field cage E 4 V / cm readout chambers High Voltage electrode ( kv) 9us readout channels Inner and Outer Containment Vessels (5 mm, CO ) beam TPC WORKING PRINCIPLE Endplates housing x x 8 MWPC Suspended field defining strips 4 V / cm ALICE TPC CHALLENGES up to x 4 charged particles in TPC 5 cm 845 < r < 466 mm drift length x 5 mm drift gas Ne, CO, N (9//5) gas volume 95 m readout pads 5 6 7

2 Challenges at high particle multiplicitiessimulated ALICE event Challenges at high particle multiplicities Can a TPC be safely operated at this high particle multiplicities and high luminosity? stability of readout chambers and field cage at high load ageing problems Can we measure with enough accuracy (tracking efficiency, p & de/dx resolution)? Cluster pile-up High granularity D High data volume Low diffusion gas (CO) Dlow drift velocity D high drift field (KV) Space charge problems (drift field distortions) Low Z gas (Ne) D little primary ionization D high gas gain (x4) The closest Drift vel. depends sensitively on temp., HV, gas composition approximation Gas gain depends sensitively on mixture 6k primaries of the Big Bang 3 September September 7 Challenges at high particle multiplicities 8 Low mass Field Cage Aluminum Can we handle the detector data throughput? FIELD FIELD CAGE CAGE DESIGN DESIGN OBJECTIVES OBJECTIVES Tedlar Fiber Pre-Pregs (pads) x (time bins) Provide Provide high high stability stability and and uniformity uniformity for: for: Honeycomb Core Nomex gas gas gain: gain: >> 44 7 Mbytes / event Pb Pb (@ Hz) Î 4 Gbyte / sec drift drift field field (4 (4 V/cm): V/cm): EErr // EEzz << -4-4 p-p temperature: temperature: (@KHz) Î 7 GByte / sec 'T 'T <<.. ooc C drift drift gas gas purity: purity: << 55 ppm ppm O O,, << ppm ppm H HO O D data compression in FEE ~ 3 cm Provide Provide high high mechanical mechanical precision precision for: for: D accurate signal preprocessing in FEE central central electrode: electrode: 5 5 mm mm readout 5 readout plane: plane: 5 mm mm 5 cm High High struc. struc. integrity integrity and and low low density density and and low low ZZ Al skins Air Tedlar CO e-identification e-identification (TRD (TRD detector) detector) minimize minimize multiple multiple scattering scattering Nomex C-fiber Ne-CO Glass fiber Total x/x ~ 3% radiation length at K = 3 September September 7 Readout Chambers Design Considerations Z (time direction): higher sampling rate Single avalanche.4 oversampling t R-I (pad direction): smaller pads limitations: # of channels (cost!) HV-GND gets critical PRF is diffusion limited cm cm temporal signal is diffusion limited 5 8 Anode wire plane without field wires FWHM = ns.6 D( t ) cm signal/noise gets critical.8 46 ALICE TPC end plate TPC Signal limitations: cm 4 gain x4 Other Other issues issues with with MWPC MWPC in in HI HI Exp. Exp. oversampling In total 557,568 pads Conclusion 63 rows with 4 x 7.5 mm (inner radius) choose the time/pad area which yields still reasonable signal (S/N > ) 3 September 7 Readout Chambers 86 cm Light composite materials for all four cylinders for a given pad area optimize aspect ratio minimize diffusion: cold gas, use high drift field 64 rows with 6 x mm 3 rows with 6 x 5 3 September 7 8 mm (outer radius) Protection Protection of of wires wires against against discharges discharges (sparks, (sparks, glove) glove) High High suppression suppression of of ion ion feedback feedback (~ (~ -5-5))

3 The Ion Tail Problem The Ion Tail Problem Ionization from 83 Kr Decay Measured with ALICE TPC prototype) 9 Aliroot: Aliroot Montecarlo data convoluted with microscopic TPC simulation g with measured p signal cluster peaks cluster peaks NA49-FTPC montecarlo amplifier data response 5 Aliroot data convoluted g with measured p signal cluster peaks NA49-FTPC amplifier response Pad Row: 9 Nr samples after zero suppression: 3 Nr clusters: 76 Mean time between clusters:. s Signal corresponding to MIP TPC FEE OVERVIEW Pre-Amplifier Shaping Amplifier (PASA) DETECTOR L: 6.5 s KHz Front End Card (8 CHANNELS) power power consumption < 4 4 mw mw // channel channel L: < s Hz Noise < 3 e MIP = 3x 4 e C d OPA C f T p T T OPA OPA < mv 3mV drift region 9 s gating grid kapton cable anode wire 8 CHIPS (6 CH / CHIP) 8 CHIPS (6 CH / CHIP) ALTRO PASA ADC Digital RAM Circuit Custom Backplane RCU integrator with continuos reset INPUTS PZ differentiator Process x T-bridge diff output amplifier Production Engineering Data AMS CMOS.35 m 573 PADS pad plane CUSTOM IC (CMOS.35 m) CUSTOM IC (CMOS.5 m ) (3 CH / RCU) Area Yield 8 mm 95% MIP = 4.8 fc S/N = 3 : DYNAMIC = 3 MIP CSA SEMI-GAUSS. SHAPER GAIN = mv / fc FWHM = 9 ns BIT MHz BASELINE CORR. TAIL CANCELL. ZERO SUPPR. MULTI-EVENT MEMORY single channel Noise Parameter Conversion gain Shaping time Non linearity Requirement < e mv / fc 9ns < % Production 56e (pf) mv / fc 88ns.% Crosstalk <.3% <.% OUTPUTS Power < mw mw / ch 5 6 ALTRO Block Diagram ALTRO layout and production data Data Processor Process HCMOS-7 (.5 µm) TSA -bit 5 MSPS 4 MSPS Correction of: Slow drifts and systematic effects Non-systematic effects Tail filtering Data compression 4-bit back linked format Channel address Time stamp Memory 5 kbyte 4 or 8 buffers Area Dimensions 64 mm mm Input Signal Transistors 6 milions Embedded memory 8 kbit Supply voltage.5 V Power 6mW / channel Package TQFP-76 L: acquisition 4-bit wide bus Bandwidth: Mbyte/s L: event freeze Trigger signals ER ( wafers) Mass prod. (5 wafers) Apr Dec Readout bus Yield 84% 7 8 9

4 ALICE TPC ReadOut chip (ALTRO) Front End Card: Layout, Cooling and Mounting ALTRO CHIP current monitoring & supervision ALTROs Shaping Amplifiers ADC Baseline Correction I Tail Cancellation Filter Baseline Correction II Zero Suppression Data Format Multi-Event Buffer COOLING PLATES (COPPER) DATA PROCESSOR voltage regulators power connector control bus connector 9 mm WATER COOLING PIPE readout bus connector GTL transceivers (back side) 55 mm INNER READOUT CHAMBER 9 Installation of last Readout Chamber (Summer 5) Installation of Front End Electronics (Feb-May 6) Commissioning the TPC ALICE COSMIC Trigger (ACORDE) Laser tests Cosmic tests Only two sectors can be instrumented at a time with Low Voltage Power Supplies Cooling Commissioning objectives Each pair of sectors continuously operated over 48 hours TPC in June 6 Test functionality of all components gas stability, noise, signal tails, gain homogeneity, space resolution, etc. 3 4

5 TPC pre-commissioning (6); some results Noise measurements First cosmic rays events σ NOISE Mean =.79 RMS.897 Outer ROC Inner ROC Y (cm) X (cm) (PASA noise + ADC noise ) =.7 ADC counts r.m.s. of the pedestals (s NOISE ) is below ADC count as required in the Technical Design Report 3-dimensional view of a cosmic muon track traversing sectors 5 6 First cosmic rays events First cosmic rays events Outer ROC Outer ROC Inner ROC Front view Side view Inner ROC 3-dimensional view of a shower induced by cosmic rays High occupancy shower induced by cosmic rays 7 8 Tail cancellation and baseline restoration cluster width diffusion coefficient (/) ALTRO Signal Processing High Multiplicity cosmic rays Occupancy ~ 5% raw samples after signal processing Z Sigma (cm).4 Z TDR value: m / cm Run 568 TPC signal shape Fit: σ = p +(z)p Outer sector : p =.936 p =.689 Inner sector : p =.768 p = Z Position (cm) width ( of a gaussian fit) of the charge signal in the longitudinal coordinate Z (drift direction) as function of Z (cut on inclination angle tan( ) <.5) 9 3

6 Signal attenuation electron attachment A.9 cm.8 5 cm.7.6 A Position resolution 5 B 4 A 3 O extracted from fit to our data [O] (ppm) Drift length (cm) Signal loss due to electron attachmnet in O At very low O content, signal loss is 3.5% / m / ppm B B [O] =.3 ±.3 ppm.5.4 Cluster charge (ADC counts) Signal Signal loss. A space point resolution in Y direction as function of z position (cut on inclination angle tan(i) <.5) B space point resolution in Y direction as function of z position (cut on inclination angle tan(i) <.5) Q total (total charge in the cluster) as function of z position (drift length) relative decrease over the full TPC length: ~% TDR specs [O] < 5ppm decrease due to electron capture and tail losses (diffusion) - ppm of O measured with oxygen analyser 3 September 7 /3 of nominal flow Expected even better performance at full flow 3 3 September 7 3 Laser tracks in the TPC Laser system for the TPC Test and Calibration Design Design testing of electronics Measured Measured alignment of ROC variations of drift velocity Drift velocity =.73 cm/ps Î agreement with drift monitor detector (Goofie) 3 September September 7 Descend to the cavern (Jan 7) 34 TPC in the cavern floor hours for descent 3 September September

7 Ready to go inside the spaceframe (Jan 7) Inside the Space-frame (May 7) ITS at TPC center Muon absorber Summary The largest TPC ever built will be at the hart of the ALICE Experiment to study the ultra-relativistic collision of heavy ions High optimization of all components and some innovative aspects Commissioning, two sectors at a time, with cosmic and laser tracks since June 6 Preliminary results show many features achieve the expected performance: Gas pressure: excellent stability! Noise < e Signal well separated from noise ( S/N > 3 for MIP) Space point resolution ~ mm after.5m of drift Back-up Slides o Jan Mar 7: installation underground in the ALICE Detector o Nov 7: start commissioning of full TPC in its final position in ALICE 39 4 LHC: The closest approximation of the Big Bang How to Measure in a High Track Density? Mini Bangs. Head-on ultra-relativistic collision of two heavy ions TPC WORKING PRINCIPLE Ionization from 83 Kr Decay Measured with ALICE TPC prototype). The energy made available in the collision generates many quarks and gluons 3. Quarks and gluons interact under the effect of strong interaction: matter tends towards equilibrium 4. The system expands and coolsdown 5. Quarks and gluons link together to form hadrons 4 4 3

8 Field Cage - Construction Field Cage - Construction Suspended Al-mylar strips Streched Al-mylar central electrode Endplates to hold ROCs Mechanical precision mm Macrolon Rods (xx8) support the strips (à la NA49) support the HV degraders distribute the gas distribute laser tracks (à la STAR) CERN - PH / TA How to Measure in a High Track Density? The ALICE Event Display Projection of the drift volume into the pad plane dn ch / dy = 8 x 4 charged particles Projection of a slice ( o in ) How to Measure in a High Track Density? TPC OCCUPANCY (*) IN THE PAD-TIME SPACE INNERMOST PAD ROW: 5% OUTERMOST PAD ROW: 7% AVERAGE OCCUPANCY: 5% (*) Occupancy = N ABOVE / N ALL Nr Pixels 573 pads x 5 time bins pad row 4% occupancy! amplitude 5 5 Occupancy figure for an ideal cancellation of the ion tail! Nr hits = ALICE CHALLENGE slice: N ch (-.5< <.5) o in = pad number time bin CLUSTER AT THE INNERMOST PAD ROW OF THE TPC Architecture and Main Components Pre-Amplifier Shaping Amplifier (PASA) Each of the 36 TPC Sectors is served by 6 Readout Partitions PASA ALTRO Impulse Response Function 8 ch 3 ON DETECTOR OFF DETECTOR DETECTOR 8 ch 8 ch 8 ch 8 ch 8 ch Front-end bus ( MB / s ) Bus controller ( conf. & R/O ) Local Monitor and Control Data Proc. and Memory BOARD Controller Control Network (I C-serial link) RCU DAQ int. (DDL-SIU) DCS int. (Ethernet) Trigger int. (TTC-RX) DDL ( MB/s ) DCS ( MB/s ) TTC optical Link (Clock, L and L ) out- FWHM = 9 ns out+ Q = 49 fc A (t / ) 4 e -4(t/ ) Overall: 4356 Front End Card 6 Readout Partitions

9 ALTRO PERFORMANCE Baseline Correction I Effective Number of Bits vs Input Frequency Baseline drift Systematic perturbation Quartz Jitter: LSB 5ps r.m.s. 9 Amplitude Uncertainty: ENOB in f jitter Fixed pedestal Systematic perturbation Slow drifts Combination - 8 ALTRO6 -.5 bits at 4.8 MHz ADS AD9 TDA8766 HI57 dbc HD HD3 7-7 HD Fin (MHz) -9 BW at PASA output f (MHz) fpd = H(z) = (-z - ) / (-.5)z Tail Cancellation Filter Baseline Correction Functions signal (ion) tail suppression pulse narrowing improves cluster separation gain equalization Architecture 3 rd order IIR filter 8-bit fixed point sc arithmetic single channel configuration 6 coefficients / channel Filter compensates undershoot Filter Narrows the pulse After Tail Cancellation Filter Characteristics: Double threshold Corrects non-systematic perturbations during the processing time Moving Average Filter (MAF) Double threshold scheme (acceptance window). Slow variations of the signal Baseline updated Operation. Fast variations of the signal Baseline value frozen BC II After Baseline Correction II A fixed threshold can now be applied safely bits 8 bits 8 bits bits word 3 rd order word extension IIR filter rounding sc sc sc sc input output bit bit Z - L Z - L Z - L3 bsl bsl calculation frozen bsl din n 8 8 din - bsl + offset, dout 3, dout < din - bsl + offset - 4, dout > 3 K K 8-bit fixed point arithmetic 5 K3 3 H( z) K z K K3 8 H( z) z z z z 8 Unsigned -bit FIR system 5 Tail cancellation and baseline restoration cluster width diffusion coefficient (/) ALTRO Signal Processing PRF width (cm) TDR values mm for short pads 3mm for long pads.5..5 Run 567 TPC cluster shape Fit: σ = Outer sector : p p +(z -z)p d =.556 =.344 p Inner sector : p =.78 p = Z position (cm) width ( of a gaussian fit) of the charge signal in r- direction (pad row) as function of Z (cut on inclination angle tan( ) <.5)

10 ALTRO PERFORMANCE Baseline Correction I Effective Number of Bits vs Input Frequency Baseline drift Systematic perturbation Quartz Jitter: LSB 5ps r.m.s. 9 Amplitude Uncertainty: ENOB in f jitter Fixed pedestal Systematic perturbation Slow drifts Combination - 8 ALTRO6 -.5 bits at 4.8 MHz ADS AD9 TDA8766 HI57 dbc HD HD3 7-7 HD Fin (MHz) -9 BW at PASA output f (MHz) fpd = H(z) = (-z - ) / (-.5)z Tail Cancellation Filter Baseline Correction Functions signal (ion) tail suppression pulse narrowing improves cluster separation gain equalization Architecture 3 rd order IIR filter 8-bit fixed point sc arithmetic single channel configuration 6 coefficients / channel Filter compensates undershoot Filter Narrows the pulse After Tail Cancellation Filter Characteristics: Double threshold Corrects non-systematic perturbations during the processing time Moving Average Filter (MAF) Double threshold scheme (acceptance window). Slow variations of the signal Baseline updated Operation. Fast variations of the signal Baseline value frozen BC II After Baseline Correction II A fixed threshold can now be applied safely bits 8 bits 8 bits bits word 3 rd order word extension IIR filter rounding sc sc sc sc input output bit bit Z - L Z - L Z - L3 bsl bsl calculation frozen bsl din n 8 8 din - bsl + offset, dout 3, dout < din - bsl + offset - 4, dout > 3 K K 8-bit fixed point arithmetic 5 K3 3 H( z) K z K K3 8 H( z) z z z z 8 Unsigned -bit FIR system 5 Tail cancellation and baseline restoration cluster width diffusion coefficient (/) ALTRO Signal Processing PRF width (cm) TDR values mm for short pads 3mm for long pads.5..5 Run 567 TPC cluster shape Fit: σ = Outer sector : p p +(z -z)p d =.556 =.344 p Inner sector : p =.78 p = Z position (cm) width ( of a gaussian fit) of the charge signal in r- direction (pad row) as function of Z (cut on inclination angle tan( ) <.5)

11 Laser tracks in the TPC Alignment of laser beams ongoing electrons emitted by the central electrode by photoelectric effect 55 7