Transcript Method 1

Particle identification by energy
loss measurement in the
NA61 (SHINE)
experiment
Magdalena Posiadala
University of Warsaw
Outline:
►
►
►
►
Energy loss measurements in NA61 experiment at CERN
Energy loss parametrisation
Methods of Particle Identification in NA61 in the region of
interest for the T2K.
Energy loss measurement in the region below 1GeV/c
 Two methods of approach to analysis:
 Method 1 – fits done using TMinuit package
 Method 2- using probability functions
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
2
NA61 (SHINE) detector
Max
72 points
Max
90 points
TPCs are filled with mixtures of Ar+ CO2 (90:10) for VTPCs and (95:5) for MTPCs
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
3
Energy loss
measurement
►
►
The optimization of the parameters required for the
determination of energy loss in the TPCs was performed
using the method developed by NA49. (For details see Eur.
Phys. J , C45 (2006) 343)
Corrections for the following effects were applied
during the calibration of the NA61 data:
 signal loss due to threshold cuts; corrections were obtained using
Monte Carlo calculations for gas mixture used in NA61,
 time dependence of the TPC gas pressure,
 residual time dependence of the measurements (day/night),
 charge absorption during the drift,
 differences in the TPC sector gain factors,
 differences in the amplification of the preamplifiers and edge
effects at sectors boundaries.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
4
Energy loss
measurement
Preliminary distributions of corrected dE/dx values for positively (white) and
negatively (yellow) charged particles with momentum chosen around 5 and 15
GeV/c.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
5
Energy loss
parametrisation
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
6
Energy loss
parametrisation
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
7
dE/dx vs number of measured
points on the track
1 island-> vtpc
2 island –>
mtpc+vtpc
Estimated resolution of dE/dx measurements of 4-5% can be achieved for
particles passing through both the vertex and the main TPC chambers.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
8
Total momentum
N>10
N>40
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
N>30
N>50
9
 dE/dx for total momentum
p< 2 GeV/c
tracks
with
high
dE/dx
errors
are
rejected
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
10
Methods of Particle Identification in
NA61 in the region of interest for the T2K
►
►
In NA61 (SHINE) experiment particle identification is
possible using ionization measurements in active
volume of the TPCs and time-of-flight measurements
in ToF counters.
For ToF-F measurements see next presentation of
Sebastien Murphy.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
11
Comparison of
NA61 data with T2K
simulation
► JNUBEAM
JNUBEAM –simulation of the T2K  beam
NA61, 2C data,
N>30
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
12
Tracks with and without TOF-F
N>30
dE/dx info
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
13
Energy loss
measurement
in the region below
1GeV/c
Statistics below
1GeV/c – NA61 Data
in %
In T2K bins
ptot:
200MeV/bin x
50 bins
till10GeV/c
:
20 rad/bin x
20bins
till 0.4 rad
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
15
Statistics for N>30,
p<1GeV/c
p (GeV/c)
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
16
Statistics below
1GeV/c -2C Data
in %
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
17


dE/dx region below
1GeV/c
Preliminary plot of the
reconstructed dE/dx
values versus momenta
(p<1.2 GeV) for
positively charged
particles, together with
the Bethe-Bloch curves
for positrons, pions,
kaons, protons and
deuterons.
 of dE/dx vs dE/dx
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
18
Smearing with Gaussian theoretical
distributions of energy loss
►
►
►
For whole tracks we can calculate theoretical energy loss dE/dxteor using
track total momentum and applied parametrisation functions.
In this way we get theoretical distribution of energy loss in momentum
bins (see example plot on the right).
Smearing with Gaussian function
Example ->
dE/dx
data
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
19
Smearing with Gaussian theoretical
distributions of energy loss (2)
The theoretical distributions of energy loss are smeared with Gaussian function
following this rules:
Mean = dE/dxtheor calculated for each track
Sigma = dE/dx for the each track; from NA61 data
After Gaussian
smearing
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
20
Methods of analysis
►
►
►
►
Having theoretical energy loss distributions smeared for
each of the particle hypotheses I start to extract raw
particle yields using 2 methods:
Method 1-> Making fits using TMinuit Package in Root
Method 2 ->Calculating probability functions for 4
particles hypotheses
For more details look at the example in the next few slides.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
21
Momentum range (0.16-0.3)
GeV/c
►
Preliminary plot of the
reconstructed dE/dx
values versus
momentum (p<1.2
GeV) for positively
charged particles,
together with the
Bethe-Bloch curves
for positrons, pions,
and only kaons in
selected momentum
range.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
22
Momentum range (0.16-0.3)
GeV/c
►
We have theoretical smeared with Gaussian energy
loss distributions (right) and NA61 data (left)
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
23
Momentum range (0.16-0.3)
GeV/c
► Method
1 -> uses 2 function for minimization
 N data  (N theor _   N theor _ K  N theor_ p  N theor_ e ) 

  


N
data


2
2
Theoretical distribution of energy loss is fitted
to data
► Calculation of the fit parameters ,,,
►
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
24
Momentum range (0.16-0.3)
GeV/c
► Method
1
► TMinuit fits
► Results
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
25
Momentum range (0.16-0.3) GeV/c
►
Method 2->Probability functions; for each of the bin i
assuming particle hypotheses I calculate probability function
defined as:
i
prob 
N
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
i
N
i
i
i
i
 NK  N e  N p
26
Momentum range (0.16-0.3) GeV/c
►
Method 2-> Multiply
probability functions
by distribution of the
energy loss for the
data in selected
momentum slice to
obtain colourful plots
at the bottom
picture;corresponding
to particle species.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
27
Momentum range (0.3-0.4)
GeV/c
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
28
Momentum range (0.3-0.4)
GeV/c
► Method
1
► TMinuit fits
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
29
Momentum range (0.3-0.4)
GeV/c
► Method
2
► Probability
functions
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
30
Momentum range (0.5-0.6)
GeV/c
protons are
seen
(yellow line)
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
31
Momentum range (0.5-0.6)
GeV/c
► Method
1
► TMinuit fits
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
32
Momentum range (0.5-0.6)
GeV/c
► Method
2
► Probability
functions
Almost the same
number of
positrons and
kaons, which is
of course not
true.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
33
Momentum range (0.6-0.7)
GeV/c
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
34
Momentum range (0.6-0.7)
GeV/c
► Method
1
► TMinuit fits
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
35
Momentum range (0.7-0.8)
GeV/c
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
36
Momentum range (0.7-0.8)
GeV/c
► Method
1
► TMinuit fits
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
37
Momentum range (0.8-0.9)
GeV/c
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
38
Momentum range (0.8-0.9)
GeV/c
► Method
1
► TMinuit fits
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
39
Momentum range (0.9-1.1)
GeV/c
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
40
Momentum range (0.9-1.1)
GeV/c
► Method
1
► TMinuit fits
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
41
Momentum range (0.9-1.1)
GeV/c
► Method
2
► Probability
functions
Method 2 at this stage of studies works well only for
momentum slices where theoretical lines do not cross
completely !!!
Needs further improvements (/K ratio parameterised
by function)
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
42
Summary
►
►
The resolution dE/dx strongly depends on number of
points per track. For tracks passing through all TPCs the
resolution at the level of 4 - 5% .
NA61 data cover acceptance in p,  variables for the T2K
experiment.
► Important
►
dE/dx analysis below 1 GeV/c region.
Presented two methods towards raw particle yields
► Method
► Method
1 –>making fits using TMinuit package
2 with probability functions still needs improvement>approaching towards max log-likelihood.
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
43
Further steps
►
►
On the right relative error vs energy loss distribution is presented.
Next step is to smear theoretical lines with Gaussians that have error
equal to relative error of dedx and see how it influences the presented
results
"Neutrino physics at accelerators", JINR, Dubna 27-30.01.2009
44