Low-momentum K/π identification for τ/c factories презентация

fTOF Detector for HIEPA
Time-of-Flight Cherenkov detector:
3-4σ K/π and K/p separation up to 2 GeV/c
Quartz Cherenkov radiators conduct light to the photodetector using total internal reflection
TOF detector from SuperB project in Italy was taken as a base

DIRC-like TOF for SuperB project
SuperB detector design
TOF endcap ring (beam view)
Trapeze sector with 14 MCP PMT
(beam view)
(side view)
DIRC-like TOF detector was designed for PID in forward region for a project of SuperB experiment.

Time-Of-Flight Technique
Time of flight for distance L, momentum p and mass m:
The difference in time of flight (K /π):
Particles’ time-of-flight as function of momentum for different particles (L = 130 cm)
/c

Likelihood based particle selector
In Likelihood based selector we construct out measured qualities variable which can be likelihood ratio defined by this equation:
Where LH1 and LH2 - likelihoods of the hypothesis (1) and (2) respectively.
In our case is particle K-meson or π-meson.
Considering probability density function (PDF) of the measurements are Gaussian, the likelihood can be written like this both for time-of-flight and number of photoelectrons (for threshold counting):
Likelihood distribution for two different observable separated apart
by 3 σ. For convenience we centered at 0 and 3 the LH1 and LH2 respectively, σ = 1.
Distribution of the Gaussian likelihood ratios RLH for different observable.

K/π separation
Due to this calculations, we need ~30ps detector resolution for 4σ K/π separation on 130 cm flight length at 2 GeV/c
Kaon/pion separation power (k) as function of particles’ momentum for different detector resolution
/c

Timing optimization for HIEPA fTOF
If we use 100 of small sectors instead of 12 big because we can operate with narrower time distribution of light collected
Time, ps
Time, ps
Distribution of photons by time of detection for different ring fractions
12 big sectors as at SuperB
100 small sectors

Less thickness
1.5 cm thickness
0.75 cm thickness
Twice narrower distribution, but also twice less light
Time, ps
Time, ps

Multilayer approach
Applying two or more layers with less thickness we increase probability of light detection and also increase time resolution
Time distribution of the photons detected, ps
Two layers of quartz optically separated one from each other
1st layer
2nd layer

Tilted geometry
Tilted geometry with an angle α could be applied to decrease the number of side reflections.
Less time of propagation spread
Less light loss caused by imperfections of the surface
Less time spread caused by Cherenkov light of δ-electrons
IP
Also few layers could be applied here as was proposed on the previous slide

Threshold counter approach
On some α the light will propagate directly to the end without reflection
For K and π with the same momentum this angle will be different because of different masses
Applying photo absorbers on every surface allows to detect the light from one particle and cuts the light from another
This solution makes us less dependent on photodetector timing
Angle α of propagation without reflections for K and π with different momentum. →
α, deg
Momentum, GeV/c

Threshold counter performance estimations
Geometrical factor for the amount of the detected photons vs distance from photodetector (cm).
It doesn’t include light yield dependent on momentum and detection efficiency.
Pion
Kaon
Performance illustration of the sector with light absorbers and the thickness of 3 cm for the 1 GeV K and π with an α of 23 degrees
There’s a region between 18 cm and 60 cm where the separation can be done
The separation below 18 cm can be performed using time-of-flight technique

Refractive index variations
Optimization of the α angle can be done by refractive index variations.
This approach allows saving a lot of space and covering wider spectrum of particles by threshold separation

TOF + Threshold
At the region near photodetector a lot of light is detected, but from both K and π.
In this case TOF can be used for separation.
Taking an average time for ~500 photons from every particle with 150 ps time resolution (sigma) gives us 7 ps sigma result.
It gives ~20σ K/π separation power on 2 GeV/c.
Pion
Kaon
Time distribution of the photons detected, ps
Pions and kaons at 2 GeV. Red smooth line is a gaussian fit of pion light time distribution.