7. Particle identification

Transcription

1 7. Particle identification in general, momentum of a particle measured in a spectrometer and another observable is used to identity the species velocity time-of-flight Cherenkov threshold transition radiation energy loss energy measurement calorimeter chap.

2 special signatures: photon - total energy in crystal or electromagnetic sampling calorimeter + information on neutrality neutron - energy in calorimeter or scintillator w. Li, B, 3He + information on neutrality muon - only de/dx in thick calorimeter, penetrates thick absorber K0, reconstruction of minv of decay products neutrino only weak interaction, charged or neutral current

3 7.1 Time of-flight τ time difference between two detectors with good time resolution start and stop - counter typically scintillator or resistive plate chamber, also calorimeter (neutrons) coincidence set-up or put all stop-signals into TDC (time-to-digital converter) with common start or stop from 'beam' or 'interaction'

4 for known distance L between start-and-stop-counter limiting case require e.g. at separation up to Cherenkov counter or scintillator for

5 difference in time-of-flight for L = 1 m but of course distance L can be larger $$ detector area for a given acceptance

6 particle indentification (PID) via time-of-flight at moderate momenta mass resolution: p = βγm with restmass m, β = L/τ use units: c = 1 =E 2 usually error in time measurement dominates

7 7.1.1 Resistive plate chambers: gas detector for precise timing measurement (material from talk by C. Williams on ALICE TOF) how to get a good timing signal from a gas detector? where is the problem? normally signal generated in vicinity of anode wire, cathode timing determined by drift of primary ionization clusters to this wire, signal consists of a series of avalanches anode cathode spread over interval of order of 1 s no way to get precision timing (subnanosecond) idea: go to parallel plate chamber (high electric field everywhere in detector clusters start to avalanche immediately cathode induced signal sum of all simultaneous avalanches but in practise this is not so... anode

8 cathode electrons avalanche according to Townsend N = No eαx anode only avalanches that traverse full gas gap will produce detectable signals - only clusters of ionisation produced close to cathode important for signal generation. Avalanche only grows large enough close to anode to produce detectable signal on pickup electrodes if set minimum gas gain at 106 (10 fc signal) and maximum gain as 108 (streamers/sparks produced above this limit), then sensitive region first 25% of gap Time jitter time to cross sensitive region ~ gap size/drift velocity So (a) only a few ionisation clusters take part in signal production (b) size matters (small is better)

9 first example: Pestov chamber (~1970) 30 years ago Y. Pestov: realized importance of size, Pestov chambers - gas gap of 100 m gives time resolution 50 ps first example of resistive plate chamber glass electrode and metal electrode cathode 100 µm gap 12 atmospheres anode Pestov glass Excellent time resolution ~ 50 ps or better! But long tail of late events Mechanical constraints (due to high pressure) Non-commercial glass no large scale detector ever built

10 how to make real life detector? (a) need very high gas gain (immediate production of signal) (b) need way of stopping growth of avalanches (otherwise streamers/sparks will occur) Answer: add boundaries that stop avalanche development. These boundaries must be invisible to the fast induced signal - external pickup electrodes sensitive to any of the avalanches from this idea the Multigap Resistive Plate Chamber was born

11 MULTIGAP RESISTIVE PLATE CHAMBER Signal electrode Cathode -10 kv (-8 kv) (-6 kv) (-4 kv) (-2 kv) Anode 0 V Signal electrode Stack of equally-spaced resistive plates with voltage applied to external surfaces (all internal plates electrically floating) Pickup electrodes on external surfaces (resistive plates transparent to fast signal) Internal plates take correct voltage - initially due to electrostatics but kept at correct voltage by flow of electrons and positive ions - feedback principle that dictates equal gain in all gas gaps

12 Internal plates electrically floating! Cathode -10 kv (-8 kv) Flow of positive ions (-6 kv) (-4 kv) Flow of electrons and negative ions (-2 kv) Anode 0 V In this example: 2 kv across each gap (same E field in each gap) since the gaps are the same size on average each plate has same flow of positive ions and electrons (from opposite sides of plate) thus zero net charge into plate. STABLE STATE

13 What happens if plate at wrong voltage for some reason? Cathode -10 kv (-8 kv) -6.5 kv(-6 kv) (-4 kv) (-2 kv) Anode 0 V Low E field low gain High E field high gain Decreased flow of electrons and increased flow of positive ions net flow of positive charge. This will move the voltage on this plate more positive than 6.5 kv (i.e. towards 6 kv) Feedback principle that automatically corrects voltages on the resistive plates stable situation is equal gains in all gas gaps

14 ALICE TOF prototypes: 10 gaps of 220 micron 1000 STRIP 10 H.V kv Uncorrected time spectrum σ = 66 ps minus 30 ps jitter of timing scintillator = 59 ps Time with respect to timing scintillators [ps] 1200 Entries/50 ps indeed one gets sub 50 ps time resolution Entries/50 ps 800 time spectrum after correction for slewing STRIP 10 H.V kv 800 σ = 53 ps minus 30 ps jitter of timing scintillator = 44 ps Time with respect to timing scintillators [ps]

15 Test of pre production strip: 120 x 7 cm2 read out plane segmented into 3.5 x 3.5 cm2 pads Efficiency [%] ADC bins Applied differential voltage [+ kv] Resolution (ps) 65 ADC bins Applied differential voltage [+ kv] pedestal ADC bins (a) Peak of charge spectra well separated from zero (b) No sign of streamers but how precise do these gaps of 250 m have to be?

16 gain not strongly dependent on gap size actually loose mechanical tolerance but why? Smaller gap Larger gap Higher electric field Lower electric field Higher Townsend coefficient higher gas gain But smaller distance for avalanche lower gas gain Lower Townsend coefficient lower gas gain But larger distance for avalanche higher gas gain with the gas mixture used (90% C2F4H2, 5% SF6, 5% isobutane) and with 250 m gap size these two effects cancel and gap can vary by ± 30 µm

17 Cross section of double stack MRPC ALICE TOF 130 mm active area 70 mm Flat cable connector Differential signal sent from strip to interface card Double stack each stack has 5 gaps (i.e. 10 gaps in total) 250 micron gap with spacers made from fishing line honeycomb panel (10 mm thick) PCB with cathode pickup pads external glass plates 0.55 mm thick internal glass plates (0.4 mm thick) PCB with anode pickup pads Mylar film (250 micron thick) 5 gas gaps of 250 micron M5 nylon screw to hold fishing line spacer connection to bring cathode signal to central read out PCB PCB with cathode pickup pads Honeycomb panel (10 mm thick) Silicon sealing compound Resistive plates off the shelf soda lime glass 400 micron internal glass 550 micron external glass Resistive coating 5 M /square

18 Hits in inner tracker TPC hits The red hits/track corresponds to a single particle (π in this case) TOF with very high granularity needed! Hits in TOF array array to cover whole ALICE barrel m2 and 100 ps time resolution Highly segmented - 160,000 channels of size 2.5 x 3.5 cm2 gas detector is only choice!

19 20º Cross section of ALICE detector 10º TOF TRD TOF ARRAY arranged as a barrel with radius of 3.7 m Divided into 18 sectors 20 10ﾼ R

20 54.6º 15.8º 3.6º º º TOF TRD Along the beam direction each sector divided into 5 modules i.e 5 x 18 = 90 modules in total 160 m2 and 160,000 channels 4600 I. P. Outer module 7.9º Intermediate module Central module 1275

21 0º 0.5º 1.6º Central module 8.5 cm cm 2.7º 3.7º 1º 2.1º 99 cm 114 cm º 4.8º 5.9º 7.4º 5.3º 4.3º 6.9º 7.9º 6.3º 70 Intermediate module 8.2º 9.3º 10.3º 11.4º 12.4º 13.4º 14.5º 15.5º 16.5º 17.5º 18.5º 19.5º 20.5º 21.5º 22.5º 23.4º 24.4º 25.4º 26.3º 27.3º 9.8º 11.9º 10.8º 8.7º 13.9º 12.9º 16º 14.9º 20.1º 18º 17º 19.1º 22.1º 21.1º 23º 23.9º 24.9º 25.9º 26.8º cm 147 cm Outer module 27.3º 28.2º 29.2º 30.1º 31.0º 31.9º 32.8º 33.7º 34.6º 35.4º 36.3º 37.1º 37.9º 38.8º 39.6º 40.4º 41.2º 42.0º 42.8º 43.5º 44.3º 27.8º 28.7º 29.6º 30.5º 31.5º 32.3º 33.3º 34.2º 34.9º 35.8º 36.6º 37.4º 38.3º 39.2º 40.1º 40.8º 173 cm cm 41.6º 42.4º 43.1º 43.9º

22 7.2 Specific energy loss use relativistic rise of de/dx - but is that possible with Landau fluctuations? effective way to suppress fluctuations: make many measurements of de/dx and truncate large energy-loss measurements (or also small) mean energy loss rel. to mimimum ionizing energy loss distribution and p > min.ion for protons normally only separation excluded

23 normally due to Landau tail very large overlap pion and kaon (e.g.) truncated mean method: vv several measurements and truncation of the % highest de/dx values for each track

24 alternative: method Likelihood method for several probability that pion produces a signal x: for each particle measurements probability for probability for kaon: measurements

25 multiple energy loss measurement in TPC (TPC/ Two gamma collaboration, LBNL 1988) in the mean time 3% have been reached (NA49 at SPS w. Ar/CH4) and in ALICE 6% (Ne/CO2/N2) (simulation)

26 de/dx performance of ALICE TPC with cosmics Particle Identification

27 7.3 Transition radiation effect: see chap.2, particles with Lorentz above about 1000 emit X-ray photon when crossing from medium with one dielectric constant into another, probability order 1% per boundary crossing mean energy of transition radiation photon as function of electron momentum energy loss distribution for 15 GeV e, in transition radiation detector

28 transition radiation detector (schematic) energy loss (excitation,ionization) plus transition radiation distribution of number of clusters above some threshold for 15 GeV e,

29 e/ separation in transition radiation detector - total energy loss - cluster counting method e/ separation at 15 GeV in a Li-foil radiator novel type: ALICE TRD, last hour if time

30 7.4 Cherenkov radiation real photons emitted when v > c/n v > c/n induced dipoles not symmetric non-vanishing dipole moment v < c/n induced dipoles symmetric, no net dipole moment AB= t c AC= t c/n cos θc = 1/ n

31 threshold effect: radiation for > l/n, asymptotic angle c = arc cos (l/ n) number of Cherenkov photons per unit pathlength in interval 1-2 (see chap.2) (Z = charge in e) in case of no dispersion (n const. in interval)

32 application of Cherenkov radiation for separation of particles with masses m1, m2 at constant momentum (say m1 < m2) to distiguish: particle 1 above threshold 1 > 1/n and particle 2 at most at threshold 2 = 1/n or in = nm range, lighter particle with 12 >>1 radiates photons per cm for radiator of length L and quantum efficiency q of photocathode and for threshold at N0 photoelectrons defines the necessary length of the radiator

33 required detector length for N0 = 10 and q = 0.25

34 , K, p separation with Cherenkov detector: use several threshold detectors p = 10 GeV/c : = / = K: p: condition no radiation: < 1/n or 1/ > n pion trigger K kaon trigger p proton trigger

35 differential Cherenkov detectors: selection of velocity interval in which then actually velocity is measured accept particles above thereshold velocity min =1/n detect light for particles between min and a value t where by total reflection light does not propagate into (air) light guide cos C= 1/n critical angle for total reflection: sin t =1/n cos t = (1-1/n2) > range: 1/n < < 1/ (n2-1) example: diamond n = > 0.41< < i.e. = 0.04 window selected if optics of read-out such that chromatic aberrations corrected -> velocity resolution / = 10-7 can be reached principle of DISC (Discriminating Cherenkov counter)

36 Ring Imaging Cherenkov counter (RICH): optics such that photons emitted under certain angle form ring of radius r at image plane where photons are detected often use spherical mirror of radius Rs projects light on to spherical detector of radius RD focal length of spherical mirror : f = Rs/2 Cherenkov ligth emitted under angle c radius of Cherenkov ring at detector : r = f c = Rs/2 c = (n cos (2r/Rs))-1 Photon detection: photomultiplier or multiwire prop. chamber or parallel plate avalanche counter filled with gas that is photosensitive, i.e. transforms photons into electrons e.g. vapor addition of TMAE ((CH3)2 N)2 C = C5 H12 N2 Eion = 5.4 ev

37 example: K/ separation at p = 200 GeV/c photons detected in MWPC filled with He (83%), methane (14%), TEA (triethylamine, 3 %) CF2 entrance window (UV tranparent)

38 Event displays CERES RICH 1 electron produces about 10 photons Johanna Stachel

39 CERES Electron Identification with TPC and RICH RICH rejection vs. e efficiency rejection via TPC de/dx Combined rejection - e.g. at 1.5 GeV/c at 67 % e eff. factor rejection Johanna Stachel

40 comparison different PID methods for K/ separation: a detector system for PID