Some results on the CBM straw TRD Monte Carlo simulation

1
Some results on the CBM straw TRD Monte Carlo
simulation

• Vladimir Tikhomirov
• P.N.Lebedev Physics Institute, Moscow
• Presented on the CBM meeting, 6-8 October 2004,
  GSI, Darmstadt, Germany

2
Motivation and method

• The main goal - to optimise general layout of
  straw TRD (straw diameter, radiator parameter)
  from the point of view of pion/electron
  separation
• Stand-alone program primary developed for the
  ATLAS TRT simulation is used
• TR simulation in so called field transport
  approach dE/dx energy loss using PAI model
  GEANT3 for geometry description and transport of
  particles
• The model has been carefully checked by
  comparison with ATLAS test beam data and has
  shown an agreement with a few percent accuracy in
  pion and electron spectra
• Rejection power was optimised for single particle
  crossing the detector. The double particle case
  is also considered

3
Field transport approach in TR simulation

• The main features of field transport approach for
  TR simulation (P.Nevski)
• Charge particle is accompanied by a TR photon
  spectrum
• The spectrum can be partially absorbed in the
  detector material
• If a medium is a radiator, an additional TR
  spectrum is generated taking into account the
  local radiator geometry and the particle
  direction
• Standard tracking is done by GEANT for the parent
  particle only
• The part of photon spectrum above the photon
  production cut-off parameter in GEANT is
  instantiated as a photon, propagated by GEANT as
  an independent particle
• The part of photon spectrum below the cut-off
  energy is not distinguishable from the particle
  itself and follow the same trajectory
• If the parent charged particle encounters an
  important perturbation like strong magnetic
  filed, scattering with a large angle, etc., the
  accompanying TR spectrum also can be instantiated
  as independent photons

4
ATLAS test beam results

• Many dedicated test- beam measurements with
  different prototypes have been done to measure
  pion/electron spectra and to compare with MC
  predictions
• Even such a tiny effects as the TR production on
  the straw walls are described by the MC model

TR radiation from straw walls has to be taken
into account to describe spectrum above 8 keV
5
ATLAS test beam results (2)

• Very good agreement between data and MC for
  different particles sort, energy and detector
  parameters and conditions

6
Geometry of the simulated detector
Straw layers

• One module consists of six identical layers
  radiatorstraws
• Straws are shifted from layer to layer by the
  straw radius
• Whole TRD - three modules, 18 layers
• Single particle crosses all the 18 straws with
  normal incident and spread of position across the
  straw radius

Radiators
Particle
7
Detector parameters

• Two options for straw diameter 4 and 6 mm
• Straw wall - 60 µm of Kapton film
• Anode wire - tungsten 30 µm in diameter
• Gas mixture - 70Xe, 20CF4, 10CO2 at normal
  conditions
• Radiator regularly spaced polypropylene films

8
Pion and electron spectra in one straw

• Example of energy loss in 4 mm straws for pions
  and electrons
• Mean energy loss is around 2.2 keV for pions and
  5.3-6.9 keV for electrons (slightly different in
  different layers due to TR accumulation effect)
  for one radiator thickness of 2.6 cm
• Around 0.35-0.5 of absorbed TR photons with mean
  energy 8-10 keV per layer

9
Pion integral spectra

• Probability to exceed threshold 6 keV in single
  straw is around 5

10
Electron integral spectra

• Electron energy loss in the straw is higher
  compared to pion loss due to produced and
  absorbed TR photons
• Probability to exceed 6 keV - 32-40 for
  different straws along the module
• TR accumulation effect is clearly seen

Straw in sixth layer
Straw in first layer
11
Cluster counting method

• Calculate the number of straws on the pion or
  electron track with energy deposition above some
  threshold around 4-9 keV (high threshold hit, or
  cluster)
• Choice the number of clusters on TRD track
  corresponding to 90 efficiency for electrons
• Calculate the fraction of pions above this number
  - pion efficiency. The inverse value is rejection
  power.

Number of clusters, corresponding to 90 of
electrons above this value
12
Optimisation of radiator parameters

• Start with 3.6 cm radiator
• The rejection power is 1.5 order of magnitude
  better than required 1
• Optimal energy threshold is around 6-7 keV
• Almost no difference between 150-250 µm gap
  between foils
• The larger gap is preferable smaller number of
  foils and less amount of material

13
Optimisation of radiator parameters (2)

• 20 µm foil practically the same rejection as
  with 15 µm foils
• 15 µm is preferable less amount of material

14
Optimisation of radiator parameters (3)

• 25 µm µm foils worse rejection compared to 15 µm
  or 20 µm

15
Optimisation of radiator parameters (4)

• 2.6 cm radiator worse rejection compared to 3.6
  cm radiator, but still is much better, than
  required 1
• 400 µm gap between foils deterioration of
  rejection power

16
Optimisation of radiator parameters (5)
17
Optimisation of radiator parameters (6)
18
Optimisation of radiator parameters (7)

• 1.6 cm radiator seems too small. Rejection power
  is around 1, but one should keep in mind that we
  still consider an ideal conditions (single
  particle case, 100 of straw efficiency, etc.)

19
Optimisation of radiator parameters (8)
20
Optimisation of radiator parameters (9)
21
Optimisation of radiator parameters (10)

• 6 mm straws the rejection is even better than
  for 4 mm
• Optimal threshold is moved to 8 keV

22
Optimisation of radiator parameters (11)
23
Optimisation of radiator parameters (12)
24
Dependence on particle energy

• All the results above are for 20 GeV particle.
  For lower energy (except of 1 GeV) the expected
  rejection is better

25
Dependence on particle energy (2)

• Deterioration of rejection power at very low
  particle energy due to small TR yield, at high
  energy - due to relativistic dE/dx energy loss in
  the straw gas for pions

26
Rejection as a function of detector efficiency

• Skip the signal from several straws to simulate
  non-operational straws in the detector
• For detector configuration with 2.6 cm radiator
  the maximum allowed number of non-operational
  straws on the particle track should not exceed
  2-3 among of 18 (fraction of operational straws
  gt80)

27
Rejection for double particle

• Rejection in high multiplicity environment
  usually much worse compared to single particle
  case. An example of such deterioration is shown
  here
• Two neighbour pions (crossed the same straws in
  all 18 layers) give two order of magnitude worse
  rejection compared to single pion

28
Possible double pions/conversions analysis
methods 3 thresholds
3 threshold analysis (1) LLT200 eV (2)
MLT?1.5-2.0 keV (3)HLT?6 keV

• Electrons and pions

• Electrons and conversion inside of TRD

N clusters on track between Thr 1 and 2
N clusters on track above Thr 3
29
Conclusions

• Monte Carlo simulation has been done to optimise
  general layout of straw TRD from the point of
  view of pion/electron separation
• TRD consisted of three modules with six layers of
  radiator/straws for each module was considered
• Different detector parameters - straw diameter,
  total radiator thickness, radiator foil thickness
  and gap - were optimised
• An optimal parameters of radiator both for 4 mm
  and 6 mm straws one radiator thickness around
  2.5-3 cm, foil thickness 15 µm, gap between foils
  200-250 µm. Optimal threshold for energy loss in
  one straw - 6-7 keV for 4 mm straws and 8 keV for
  6 mm straws.
• The expected rejection power for 20 GeV single
  particle is around 0.15 for 20 GeV and
  0.05-0.08 for lower energy 2-10 GeV

30
Conclusions (2)

• The influence of different factors on rejection
  power was considered energy of particle, double
  hits and efficiency of detector
• The main conclusion it is not a big problem to
  reach the rejection power of 1 for single
  particle. The main problem is - how it will work
  in high multiplicity environment, including
  double hits and conversions? One possible
  solution is to use three-threshold analysis.