ORS'CHU .a ZEN .TRU JÜLI H G aH
Institut für Kernphysik
Response Function of the
Trigger Scintillation Detector
Pawel Moskal
Berichte des Forschungszentrums Jülich ; 2825
ISBN 0944-2952
Institut für Kernphysik Jül-2825
Zu beziehen durch' Forschungszentrum Jülich GmbH Zentralbibliothek
D--52425 Jülich Bundesrepublik Deutschland
Telex : 838556-70 kfa d
Telefon : 02461/61-6102
Telefax' 0 2461161-6103
Response Function of the
Trigger
i
Paw&l Moskal
Contents
7
Introduetion
The detection modale operation
1 .1 1V1echanism of the Iight pulse production in a plastic scintillator
,
1 .2 The light collection process
1 .3 Photomultiplier
1 .3 .1 The photoernission and electron collection processes
1 .3 .2 The electron rrtatltipic~atiort 2 The photomultiplier
.
2 .1 Plateau . .
2 .2 Mensitrernent of the single electron response 2 .3 The gain variation
3
11
12
15
17
18
20
23
23
24
27
Methods of the detector examination
3 .1 Experirnental set-up used in rneasurei lents with the "S T
source
3 .2 The experirnental set-up used in Ute measurement at a proton and pion beam
.
.
3 .3 Hit position deterrnination .
3 .4 Time walk correction 29
3 .5 Putide identification 38
5
30
32
34
35
4
The response of the detection rnodule
4 .I The detector efficiency
4 .2 A pulse charge variation
4 .3 Estirnation of the light collection efficiency
4 .4 The light pulse velocity
4 .5 Time resol ution
39
39
41
43
45
47
Acknowledgements
50
References
52
6
INTRODUCTION
7
Introduction
Thorough investigations carried out over many years revealed, that the
hadronic matter is certainly mach richer than the simple constituent quark
model expectation with only three quark and quark- antiquark hadrons.
The QCD theory predicts the existence of hadronic objects, other than the
known baryons and mesons, like glueballs and hybrids (Mt) . Presently
available data, however, does not provide sufficient information to judge,
for instance, whether f0 (975) and a0 (980) mesons reveal qq, gggq or s" . {
molecular structure . Their small two-photon partial width T ~„ and a very
narrow füll width [2 still remain unexplained . Further, there exists so .e
recent theoretical predictions that ss quarks constitute only 50% of the
0(1020) meson, which is generally regarded to possess pure s structure.
However, so far this is not confirmed experimentally . Equally interesting
is a discussion of the gluon content of eg . the 17 1 (958) meson [1_ . A few
attempts were made to predict theoretical glueball masses, but the results
are model dependant [2 : . Thus, investigations which can prove the existence of the exotic quark states are presently of significant interest in
medium energy physics.
In order to find an answer to sonne of the above mentioned problems
the detection systern, sketched in Fig . 0 .1, has heen designed . Soon it is
to be installed in the cooler synchrotron ring (COSY) at Jiilich, which will
8
INTROD UCTICN
provide a high quality proton beam with proton momenta up to 3,3 GeV/c.
Experiments are planned to allow the study of the structure of resonances
in pp -4 pp ° and pp pnM + reactions with meson masses up to the
meson «1020) and also investigations an meson - meson or rraes®r nucleon final state interaction via reactions of the type pp --> ppM1 M2 .
The aim of this work is to test the response of a scintillation detector
(indicated in Fig . 0 .1 as S1 to ionizing particles . This counter, consisting
of sixteen detection modules, will serve as a trigger of the whole detection
system. Thus the time resolution as well as a signal amplitude variation
with respect to a hit position is of a special interest . The former because
this detector will be used as a statt counter for the time of flight measurement, the atter as it will provide energy loss measurements of the
particles.
The present work is divided into two parts . In the first one the main
stages of a signal production by scintillation Counters are considered . In
the second one the first chapter presents measurements of the characteristics of the photomultiplier, whereas the second one contains a description
of the experimental set-ups as well as the method of data eva .luation . The
final chapter in turn presents the main characteristics of the considered
detector.
9
1'NTRODUCTION
Figure 0 .1 Schematic drawing of the COSY - 11 (E-5) experimental set-up.
The trajectories shown are for the p p -4 p p K+ K- reaction,
at a beam uiomenturn 2 MeV above a threshold.
drift ehambers with an active area of 43 x 168 cm 2 ,
D1,D2
scintillation counters with an active area of,
S1,S2,S3,S4
45 x 160 cm2 , 46 x 21 .6cm2 , 150 x 200 cm2, 6x 110 cm 2 , respectively,
silicon pad array with an active area of 5 .8 x 105 cm 2,
Si
monitor detector (silicon pad array - 5 .8 x 21 cm 2 ),
M
claster target,
"
T
(Courtesy M . Rook and D . Grzonka)
z
1
Figure 0 .2 Schematic diagram of a separate modale of the Sl detector.
1
2a
3
4
5
6
7
8
scintillator BC - 404 of dimensions 450 x 100 x 4 mm',
twisted strip Iight quide made of GS233 Röhm plexiglass,
cylindrical Iight guide,
photomultiplier EM' 9954B,
photorrmltiplier fase (voltage divider network) type 4244,
capton feil,
mu metal,
plastic housing,
iron protective shield
The detection module operation
There exists a certain kind of material, scintillator, which emits flashes
of light when exposed to radiation . This peculiarity enables to make use
of such materials in particle and radiation detection . The very early technique, first applied by Crooks in 1903, was based an the observation of
light flashes by means of a microscope in a darkened room . This, however, was exeedingly laborious and thus not very popular . In 1944 the
photomultiplier replaced the human eye making very efficient and reliable
particle detection possible . Since that time for allmost flfty years scintillation Counters have been successfully used in a very wide Tange of nuclear
and particle physics experiments.
The basic processes involved in the particle detection are as follows :
A charged particle, when hitting the scintillator, deposits therein a Part
of its energy, through the ionization or excitation of the electrons to the
higher molecular levels . Subsequent deexcitation processes result in photon production . A fraction of emitted photons travels to the edge of the
scintillator, from where they are transported via the light guide to the
photomultiplier . The photomultiplier first converts the light into a weak
photoelectron current, and afterwards amplifies it, so that the output signal can be analyzed by the electronics System.
11
I2
C.tIAPTER 1 . 'TUE I)E'I I,C.TIO~ A1ODUL T OPERATION
1.1 Mechanism of the light pulse production in a
plastic scintillator
Table 1 . Sonne physical properties of BC-404 plastic scintillator 3]
Density Refractive
Light output
Wavelength
Base
(% Anthracene) of rnaxirr um
[g/cm ]
index
emission [nm]
L032
1 .581
68
408
PVT
I3ul light
Rise
Pulse width Decay constant
Softening
time
F H
rnain corr poncnt attcmuition
point
[ns]
[cm]
'c]
[ns]
[ns]
2 .2
L8
120
0 .7
75
'c
The fluorescence process in organic scintillators arises frorn radiative
transitions in a single molecule a.nd does not depend on the physical state
of the material . In such a molecule the valence electron orbit als combine
to form rr - electron energy levels . The radiative transitions anong thern
account for the scintillations . The 7r electron system is schenatica.lly
shown in. Fig . 1 .1 . Dashed lines indicate vibrational sub-levels superimposed on Bach of the electro-rnagnetic levels.
0 .15e V
3
4eV
t
s,rt
s, s.
Figure 1 .1
7r
Electronic energy levels of an organic molecule.
T2, excited
So, ground state . SL , S2 , excited singlet states .
triplet states . Soo, So ' m . . etc . vitrrationaI sub-levels 4]
A
PTER
1.3
1 . TUE DETECTION MODUI,E OPERATION
BC - 404 is a ternary plastic seintinators Ternary rneans that two
admixtures are dissolved in the base material . The base of this scintillator
is polyvinyltoluene . lt seintillates as pure material, hart its light output is
relatively low . This is considerably improved by dissolving therein a small
amonnt (a few percent) of an aromatic compound called prirnary solute.
Ilowever, such a solution can not be used in Karge quantities, becailse
of a strong light absorption . Second additive (secomlary solute) allows
to get rid of this disadvantage by shifting the emitted light to a longer
wavelength.
In ternary scintillators the concentration of the admixtures is so small,
that their excitation by an ionizing particle can be negleeted . Thus, at
ferst, the particle deposit its energy in the base . Next this energy may be
transferred to the solute, hoth radiatively and non-radiatively . Ilowever,
because of the very Iow concentration of the secondary solute, this transfer
occurs rnostly between the base molecule anal prirnary admixture . The
additive is chosen, so that the energy of its first excited level is lo ser than
the energy of the first excited level of the base rnolecule . (see Fig . 1 .2)
x
x
s3x
S2
sox
y
Z
PRImARY
EXCITATION
EMISSION
Eigare 1 .2 The scintillation process in a ternary organic solution scintillator.
(X solvent(base), primary solute, Z .,s- secondary solitte) . Continuous
lines - radiative transitions, dashed lines - non-radiative transitions
Further , the excitation energy from the prirnary solute is transferred
to the secondary solute called wavelength shifter . A proper choice of the
14
CRAPTER 1 . TIIE DETECTION MODULE OPERATION
latter allows to match the seintillator ernission spectrurn to the sensitiveness of the photomultiplier . The energy of the photons emitted by the
secondary solute is very k)w in comparison with the first excited state in
polyvinyltoluene . Thus, these photons are not alle to excite the molecules
of the bare . This makes the ternary solution transparent to its own radiation, and allows to use samples of lange dimensions . The light ernission
spectrum of the BC-404 scintillator is shown in Figure 1 .3.
100 r---
0 40
4k 440 40 40 ,50
WAVELENGTH (nm)
Figure 1 .3 Norrnalized emission spectrurn
of the BC-404 scintillator [31
The radiationless processes such as the intcrnal conversion fror the
higher excited states of the solvent, or mm-radiative interinolecnlar energy transfer are ntuch faster dien the radiative processes [5] . For tha.t
reason the Signal sha .pe is determined by the light emission frorn a wavelength shifter and by the radiative energy transfers . M . Moszynski and
H . Bengston [6] proposed the foElowing equation to describe the light pulse
shape fromm the ternary scintillator
e--1/n)
(c- t f r
fG (t,
z(t)
(1 .1)
where i(t) is the light intensity, fc(t) is a Gaussian function descrihing the
rate of the energy transfer fro' a detected particle to the prirna .ry solute,
this term reflects the final rise time of the Signal . Terrn e - ' /T« lescribes
the energy transfer to the wavelength shifter and e -II «kscribes the final
15
CHAPTER 1 . THE DETECTION MODULE OPERATION
light emission . Therefore the pulse is characterized by four parameters
which can be found by fitting the equation to the measured sig/1, rr,
nal . Of course the photomultiplier response has to be taken into account.
Ho euer, equation 1 .1 may be applied only to small samples, where the
light pulse shape depends only an the composition of the seintillator . In
large scintillators the pulse shape is alfected by the light collection, the
seif-absorbtion and re-emision proeesses, Their influence manifests mainly
in shifting the light to the longer wavelength and in increasing the d.ecay
time constant . Thus equation 1 .1 is modified as follows:
-t/s
is(t) e
i(t)
-~ ~
(1 .2)
where i s (t) is the light pulse from a small sample (eq. 1 .1), e-t/6 is associated with photon transit time in a seintillator, and the term e' /'r accouts
for the seif-absorbtion and re-emission process [7].
2
L2 The light collection proeess
Physical properties of GS233 Röhm plexiglass 81
Talale 2 .
Refractive Bu.ik light attenuation Density
Material
length [cm]
index
[g/ cm3]
1 .18
600
Polymethylmethacrylat
1 .491
In a scintillator the photon emission is isotropic in all directions . For
photon angles above the Brewster angle all photons are totally reflected.
The rernainder leaves the scintillator or is partially reflected according to
the Fresnel's rules . The Brewster angle is given by
OB
=
arc sirr(
)
scirit
with riscirst - refrective index of the scintillator and
of the surrounding medium .
n o,,t -
refrective index
16
CIM
1.
DETECTION AIODIJF,F OPERATION
In a symmetrical scintillator the totally reflected photon rnay reach hach
surface, provided that it is not absorbed within the material . Thus, connecting a Chosen scintillator edge to another body of comparable refiective
index orte can guide the light to the required platte . Arnong the presently
used light guides a so called twisted strips one (see Fig . 0 .2) possesses
optirnum light collection property [91 . lt consists of separate strips glued
an to the edge of Ute scintillator and twisted 90 0 so timt they can fit the
cylindrical photornultiplier shape.
Two practical rules allow to minirnize the photons loss . Firstly, the
cross sectional area of the light guide should never &uhasch lf it is the
case then only a fraction of light, equals to the ratio of output to the
Input area, can be transferred . Secondly, a band radius should be at least
ten Limes larger than the thickness of the body . Equally important is the
light guide transmission of the fluorescence light frorn the scintillator . The
transmission of the GS233 plexiglass together with the BC-404 scintillator
emission spectrum is shown below .
ton
8o
continuous line
transmission spectrum
o
4o
dashed line
BC-404 emission spectrum
2
3() ;71
400
500
WAVELENGTH (nm)
60
Figure 1 .4 Transmission spectrum of the GS233 plexiglass [10]
versus emission spectrum of the BC-404 scintillator
The whole scintillator modale ha.s to be enclosed in a light-tight envelope . The suitahle covering may improve the collection property . However,
the best would be to leave a laycr of air surrounding the scintillator in or-
CHAPTER I . TIIE DETECT.ION MODULE OPERATION
[7
der to maxirnize the total internal reflections . In the present (-Ase the
scintillator and the light guides are wrappccl in aluminium foil, whose re-
fiectivity for the light emitted by 13C-404 material amonnts to 90% (see
Fig. 1 .5) . Takirig into account that on the a.verage photons make 10 to
20 reflections before reaching the photocathode one can roughly estirnate
that only a fcw per cent of the photons, collected at the photocathode,
originte frorn the external refkrtions.
loo
rnntinunus lin
6o
aluminimn refiectivity
1=
w
40-
dashed line
BG-404 emission spectrum
w
2o-
300
400
500
WAVELENGTH (nm)
600
Figure 1 .5 The alurnirdum rellectivity [131
Thns prhnarily three factors influence the light collection efficiency.
These are : Ute total iriternal reflection angle, the qnality of the surfaces
and the bulk light attemiation length, where the last is defined as the
length after which the nnmber of photons is diminished hy a factor r -t .
The longer the scintillator, the larger are the variations in the signal
lieight over its volnme . However, froren this fad one can make a rongh
determ.ination of the hit Position of the detected particle.
L3 Photornultiplier
Talale 3. Characteristics of the 9054E EMI photomultiplier 1111
Photocathode QE% Dynode Transit time Typica.l gain
[nsi
type
peak material
41
6 106
1)ialkali
26
BeCu
1
CHAFTER I . Tli13 DETECTION A1ODU1iE OPEI'MTI'ON
The photomultiplier serves as a converter of the weak seintillation pulse
into an electrical signal . Its operation may be considered as three subsequent processes . The photort convertion into a weak photoelecCron current,
the collection of the photoelectrorts on the firnt rnultiplier dynode and the
signal amplification by the multiplier structure.
1 .3 .1
The photoemission and electron collection proeesses
The incident light is partially absorbed in the photosensitive material
deposited on the inside of the photomultiplier glass window . The captured
photon transfers its entire energy to an encountered electron . Travelling
to the boumlary this electron will suffer an energy hiss due to electron
electron collisions . In order to leave the photocathode it has to reach
the surface with suffizient energy to overcome the potential barrier, which
always exists at any interface between material and vacuum . The finite
potential barrier imposes a minimum energy on the incoming photons.
This is reflected in the Jong wavelenght cutoff of the photocathode efficiency (sec Fig . 1 .6) . The cut in the skort wavelenght is caused by the
window material.
3O T--
1
contimious line
photocathode Q .E.
dashed line
BC-404 emission spectrurn
300 400 500 600 700
WAVELENGTH (nm)
Figure 1 .6 Quantum efliciency of a Bialkali photocathode [ll]
versus BC-404 emission speetrurn
For some semiconductor material the "work finretion" is as low as
1 .5 eV . The energy of the fluorescence light is about 3 - 4 eV (see Fig. 11) .
(,IPA1'I t I. I . TUT I)1 i °,E,1i(N ' ()1) T L! (1I'1`,i ./t 11f) i
69
Therefore, the depth at which electrons rr a.y cifigina,te and still esc.a.p}e rom
the material is about 25 nm [121 . A. photoca.thocle of this thickness rr ay
absorb no more than half of the incident li~ ;ltt.
A nirrrrber of large significarrce in scintillation connting is the photocathode gnantum eficien€cy . lt is €lehned as the ratio of the nfrrrtlrer of
photoelectrons prodnre€I to the nurr ber of incident photons . The gnanturn
ef icien€°y is a function, which is strongly dependeut on the light energy
(see Fig. 1 .6).
After ernission frone the photocathode the nlectrons travrl, in a slritaI>ly
designed electric field, towards the first dynode of the n nltiplier section.
The time of Rigid of Irhotoelectrons depends on the p€ girrt in the photocat.hode, in which they originated (see Fig. 1 .7) . These differences are the
main source of the spread of a transit time throngh the f$hotornfrltiplier.
Their are rninimized, to sorne extent, hy la.rge €lista.rrce and very high voltage between the photocathocle and the first dynode in comparison \<<ith
interdynode distances and voltaltes, respectively.
Scn if .rSp real.
lehr tocathod
"l'ypicri) (ylintnelec(rcrn
Lrn .jectories
'Njfirmt
ligh
Pliolßr.1tltnr e
In dynock N.
electron nptics
1?lcctron
rnultißl€er
a
Figure 1 .7 Basic elements of photomuliiplier tube Front [12j)
1 - 12 -- Dynodes, 13 - Anode,
14 * Focusing electrodes, 15 - Photocathode
2o
CIIArF 'EU 1 . r l ' 1!I nW '! 'a `' ! ' Iorv AfOF}t J !d
I > '.l?ATI ON
f
1 .3 .2 The electron multiplication
The first multiplier dynode is held at a positive potential of several
hundered volts with respect to the photocathode . Thus the electron impinging onto it may create in the order of one hundered new electrorts.
However, only a few of thern manage to escape.
765-
160 2Ö0 300 400 500 600 700
NTERDYNODE VOLTALE (V)
Figure L8 Serondary emission coe eiert versus energy
of the primary electron for BeCu material [13]
The secondary electrons, leaving the first dynode, are guided by an
electrostatic field to the second one, where the process is repeated . The
overall photomultiplier gain depends essentially an the niurrber of dynodes
and the secondary emission coeicient.
- b~
h rg
,
(1 .4)
where 5 - secondary emission factor, with typical 5 value of 5,
n numher of dynodes.
The b is determined by the hncident electron energy, which is proportional to the potential difference between the dynodes.
kaV
d
tS-
Assrrrning that the voltalte is equally divided among the dynodes, the
GHAPTER 1 . THE DETEGTION MODULE OPERATION
21
overall gain is then given by:
with V heing the overall supplied voltage.
The expression
dG
dV
= n V
derived from equation 1 .5 Shows the gain variation with respect to the
applied voltage . One can see that this is a very strong dependence . A one
per cent change in voltage may causa more than a ten per cent change in
gain depending on the number of dynodes.
The signal outgoing from the photomultiplier, may be regarded as a
superposition of many signals caused by single photoelectron incident on
to the first dynode . The distribution of these electrons is described by
equation 1 .2 . Therefore, the outgoing signal is an integral convolution of
a scintillator pulse and a photomultiplier response to the single electron.
I (t) _ i(t
(1
t') SER(t ') dt`
.7)
where SER(t) - the single electron response
i(t) the light pulse
One of the important quantities characterizing the photomultiplier is
the gain variance . In a simple m .inded model the production of the secondary electrons at a dynode can be assumed to be governed by a Poisson
distributione Hence, the number of electrons created on the first dynode
has a mean value , a Standard deviation V rä' and a relative variance 1/b,
and so on . lt could be pointed out, from the properties of the Poisson
statistic, that the relative variance of the mean number of electrons col-
22
CHAPTER THE DETECTION MODULE OPERATION
lected at the anode is represented by the following expression : [12]
vo2
1
1 +_ 1
1
(L8)
(5
6n
G
6 62 1
Thus, the spread in the pulse arnplitmie is dorninated by the 6 fluctuation
fror the first dynode . This is the hext reason why higher voltarge is applied
between the photocathode and the first dynode than between the following
dynodes .
Chapter 2
The photomultipIier
21 Plateau
.ln almost all experiments with scintillation cauntcrs the PI\.4 signal is
followed by a discrirninator . the scintillator is irradiated, arid the signal
alter the discrirninator is Input into a . scaler, one can observe the connting
rate (lange with the applied voltage for a specific discrirninator threshold.
The higher the supplied voltage, the larger the connting rate . However,
one or rnore plateaus can be ohserved according to the type of the source.
tourst s
Iow voltage
Pulse heigM
Gain
intermediate voltage
Figure 2A Steps Ionding
to the scintillator counting
plateau, taken from [13]
high voltage
23
Ctt .i%V'I lt 2 .
21
"1'ft
i'tl( '1 `c1 r1 C1I,' '!f'l .l ;id
The left siele of Fig . 2 .1 shows a Sketch of a 2®7 Bi spectrnm . At a
very .low voltage the entire spectrurn lies below the threshold value and
no counts are observed . Title increasing voltage rnore and more parts of
the energy spectrunr pass the threshold arni are counted . Figure 2 .2 shows
the rneasured Plateau for the tested cletector.
10 a
D10
10
1 400
1
600 1 800 2000 ' 2200
VOL TAGE (v)
Figure 2 .2 Plateou curve
rneasured using a 2' Bi
source
Set-up used in
the rneasurement
2 .2 Measurement of tim single electron response
If only one electron hits the First dynode then the electric signal collected an the anode is called the single electron response . L~y n .easuring
tlre charge in such signal one can deduce the photonzultiplier gain . At
room temperature typically thonsancl electrons unay escape froin the photocathode per second (dark current) This suggest that it would be enough
to measure a photomultiplier noise Charge spectruni, wind' should consist
of one big peak from one inciclent electron ancl further inuch smaller peaks
due to two or more electrons. However this is not the case . tone is not
alle to distinguislz any peak in such a spectran . Unfortunately Liiere is
too large noise caused by the multiplier section . The only way to get rid
of this is to trigger the electronics when a real signal is expected, even if it
CIIAPFEH
2. TIII
25
P!IOI`t)!'!t lI'1'f1'1Ji1 1~
is only one photoelectron . This can be performed l)y tneans of very weck
light, produced by the light emitting diode [14] .
etip eabte
sOk
1,ED connection
u
Ions C7;
The signal from the generator registered by the
digital oscilloscope
Figure 2 .3 Block diagram of the electronic system used
in the single electron response measurement.
ADC - analog (charge) to digital converter
The very skort pulse from a generator switches an the diode and simultaneously triggers the eharge to digital converter . The signal reflected
from the open end of a Clip cable switches the diode off . Tlra.rs the nein-er of photons in one pulse can be established by the amplitude of the
generator signal as well as by the Clip cable lengtlr . The signal from the
photomultiplier firnt is arnplified and then analizecl by the Charge to digital
converter .
CHAPTER 2 . '11I1? 1)1107«)MULTIPLUm.
26
In the case of the absence of tue diode signal one observes in the charge
spectrum a sha.rp peak raused by the noise of the amplifier . If a a.
signal
is inserted into the diode pnxlncing only a a.
photons, a a.
peak
appears in
charge spectrum.
900
w
z
z
THE SINGLE-E! FC.; ; R0N
sPEcrRum
600
w
z 300
0
oo
200
ADC CHANNEL
Figure 2 .4
'ehe single electron spectrum
Figure 2 .4 shows a rneasnrement with stach a lo‘‘«liode voltage that
only a single electron spedrum was observed, wherras in Fig_ 2 .5 the
voltage was increased and a clear two-eleetrons peak is seen.
600
w
z
z
1
o400
w
Figure 2 .5 Single and two-electrons peaks
As expected, the average charge in the Signals resulting frone two photo-
GRAFTIM 2 . 'I'TIE 1)110'1«)A41J1,73nIel? .
27
electrons is two tirnes larger than the charge in the single electron signal.
Actually, the spectrurn shown in Figure 2 .5 is a superposition of a few
peaks, however only the first two are clearly visible, the others have too
large width to be distinguishable.
2 .3 The gain variation
The nurnher ofsecondary electrons ernitted frorn the dynode depends on
the energy of the eleetron irnpinging onto it . ' his energy is proportional to
the inter-dynode potential . For that reason the gain of a photornultiplier
reveals a dependence an the overall applied voltage . Measurements of
---< 10 7
0
10 1700
6
1800 1900 2000 2100 2200
VOLTALE (V)
Figure 2 .6 Single eleetron peaks
measured for various supplied
voltages
Figure 2 .7 Photornultiplier gain
versus supplied voltage
the single electron charge spectrurn for different voltages allow to find this
dependence quantitatively . To cach spectrurn a Gaussian function was
fitted . The extracted rnean valne divided by the signal amplification gives
directly the PM gain.
The gain variation wich respect to the supplied voltage is shown in
Fig. 2 .7, The experimental points viere fitted by a polynomial of second
28
CHAPTER 2 . 'FITE PHOTOMULTIPLIER.
degree and the following coefficients were evaluated:
Log(Gain) = -2 .7 • 10
-6 . 172
where voltage is expressed in Volts .
+
1 .4 10 -2 ° V - 9 .4
(2.1)
IN/lethods
the detector
examination
This chapter contains the description of different experimental arrangements used for n easurements of the detector response . Methods of data
evaluation are presented as well.
The output signal of the detector contains information concerning the
registered particle . Usuaily the energy, the time when the particle passes
through the detector and the hit position are of rain interest . These
quantities can be determined, provided the relations between therr and the
relevant characteristics of the pulse are known . These relations are referred
to as the detector response . A method to find the specific detector response
is to measure the charge and the time spectrum due to monoenergetic
pa.rticles bomharding the detector at a fixed position . In large rnodules,
as in the present case, the response firnction charnges drastically over the
detector volume . Thus, it is necessary to scan the whole detection area.
In order to exarnine the detection module two measuremants were performed . In the frst one a colli .ma.ted "Sr electron source was used, and
the second was carried out at a monoenergetic proton/pion test beam of
2 GeV/c momentum .
29
30
CIM 'ER 3 . M .ETilODS OF TIIE 1 .MI11,C7«)R EXAM1NATION
3 .1 Experimental set-up used in measuremeuts with
the 90Sr source
The response of a scintillation counter of dirnensions 450 x 100 x 4 mm'
was exarnined . For the measurements an additional scintillation detector
(see Fig . 3 .2) was used, which served as a start counter and analyzer of the
electron energy . An electron, in order to he registered in the present electronic logic, has to have enough energy to pass through the tested detector
and to produce a signal, in a start courtter, higher than the threshold in
the discrirnirtator-2 and srnaller than the threshold in the discrirnina,tor-I,
see the electronic diagram in Fig . 3 .2 . In this manner electrons fro' a nat . row energy strip were selected frorn a contimlous ß spectrum . This allows
- cm:31o
---------------- ------------ ---z 0 .008
5
5
5
5
5
5
5
ai 0
5
5
5
5
t1
.006
__
5
5
e
5
45
4
1
THR HOLD 2
5
5
5
5
1
.1
tHR
NJ)
,
5
0 .0
0 .4
0 .8
1 .2
1 .6
ENERGY (MeV
2 .0
Figure 3 .1 The emission speetrum of a g0 Sr
2 .4
souree
to study the relative pulse charge changes along the scintillator surface.
After leading edge discrimination the tirning of the pulses is arranged
as shown in Fig . 3 .2 . The signal frorn the start detector is delayed so muck
that its leading edge cornes as the last orte, laut has an timing overlap with
the broader signals fron the tested modale . Therefore it is correlated, in
time, with a coinc,idence pulse . Such an adjustrnent provide that the TDC
unit is always started at the sarne time with respect to the hit moment.
The simn1taneous registration of charge and timing of a signal enables
CHAPTER 3. METIIODS OF TUE DE-TEC7«)R EXAMINATION
to reduce shifts in the time spectra cai .ised by the "time walk" of the
discriminators (see section 3 .4).
s .)-cE
JE'STE.'1J
C[JUNTLR
F, V
START
[11
f7 U N 1'.:
,
r
M
PM
"7. PLiTTEFF
217:CR .Zi
CUfNCI
Te,
SPI_
1 T 1'
P
UNI T
DISCR .
EP
D
A
t
A
GATE
GE NERAT J'
Figure 3 .2 The experimental set-up together with the block diagram
of the eheetrank system.
ADC - analog (charge) to digital converter,
TDC - time to digital converter
32
CIIAI'TP,'l? 3 . METIIODS O! I`Ifl 13i .:TIX :7«)i? F,'XAMINATION
3 .2 The experimental set-up used in the measurement at a proton and pinn beam
In this experiment the response of the detection rnodule was tested with
minirnnm ionizing pious and pmtons of the tost ; bearn TI 1 at CERN.
The event was accepted if a particle passed through the start and Ute
stop connter . The hit Position was defined by rneans of two delay wire
chambers (Dl,D2) with an active area of 10 cm x 10 cm . The initial particle
mornmitum was 2 GeV/c +20 MeV/c . The difference in time of fliglit
between protons and pions eqnals to 4 .36 ns, für the distance between the
start arid stop connter . This allowed to separate pions taracl protons by a
time of ffight measnrement .
STOP
S T ART
Dl
12,8 m
Fignre 3 .3 Schematic view of the experimental arrangernent
D1,D2 - deiny wire eitarnhers
START - 4ern x4cm seintillation connter
STOP -- 10 cm x 10 cm seintillation connter
CHAPTER 3. A1PITHODS OP"I'IIP 1)I,;7TG7«)11VXAAIINATION
33
The heam was inciderrt normal to the tested connter wich an angular
acceptance of -6mrad and mrad for horizontal and vertical direction, respectively . During the experiment the position of the detector was
changed a few times in ordPr to investigate the entire scintillator area.
Für each photomultiplier output, the integrated cliarge of the pulse and
the time of the pulse arrival were registerecl . The latter was measured in
reference to the common thning det(-rminod
the stop connter.
DIS'- P
T
DC
DISCR
STAPT
SPLITTER
DISCR.
STOP
SPL.fTTPR
D SCR .
'SPLITTER
D1SCR.
P T TEP
DISCR.
c)m
MR
DELAY
Figure 3 .4 Block diagrarn of the electronic systern
CIIA PTE1?. 3 . A1ET1I0D5 (J1 ' TIIE DETP«YI01?, EXAMINA'170N
34
3 .3
Hit position determination
Each delay wire chamber provides Tour timing signals (t u , t down , t left ,
trjqht), which carry information of the impact position of tim partiele [15].
The time differences
aml tlrfi - treqj, deine vertical and
zontal coordinates of the bit )oint relative to Ute center of the Aamber,
rcsp ,ectively.
rirsat,n
hori-
Figure 3 .5 The position determination by the delny wire charnber
In one of the measurements the scintillator was adjusted such that it
was covPred by part of a beam only (see Fig . 3.6) . This allowed to defim ,
the. position of the tested detector wich respect to the chairihers.
pixn
Figure 3 .6 The detector position determination
The active area of the ohamber was offline divided into elongated Pixels
parallel to the scintillator side (see Fig 3M) . For of thern the number
of events were counted for both : i) if a coincidence between chamber
and tested module was observed and ii) if chambers independently of the
tested module had a signal . Further, the ratio of case i) to case ii) was
CI7A1)7'ER 3 . MP:7'1101)S OF 771E DETP2( 7«)l?. EX A AHNT ION
3
extracted . Ideally unity was expectecl for Pixels corresponding with the
area of the tested detector, and Zero for the Test.
1 .0
0
0
0 .2
0'0 40
50
60
70
80
Y (mm)
90
10o
Figure 3 .7 Upper scintillator side in relation to the ehamber D2
Thus, the scintillator edge is defined wit ;h, an accuracy of + 1 min (see
Fig. 3 .7) . The shifts of the detector during the experiment were ( 1one with
the same precision . As a result the overa.ll resolution of the hit- position
determination amoimts to + 2 mm (FWEIM) . The fast that an efficiency
of only is observed can Im explained by accidental coincidences of
the chambers.
3,4 Time walk correction
An analog output pulse frorn a cletector is converted into a logic signal
by a discriminator . The most comrnon niethod for deriving a timing signal,
ernployed also in tie present aase, is Imding sage triggering (Fig . 3 .8) . The
logic signaI, in this technique, is generated at the mornent the analog pulse
crosses the threshold (Fig . 3 .8) . As one can sec, out of two pulses having
the same riss time, the higher crosses the threshohl earlier . Therefore,
even exactly coincident signals rnay trigger the discriminator at different
tirnes . This is refer to as a time walk effeet .
36
e)IF
CHAPTER 3 . METIIODS
L_.,
TUT
i
Output B
"Walk'
t,
.
Figure 3 .8 Walk in a discrirrminator, taken fron [Ui
The variation in the anaplitude of the inraming pulses is not the only
source of the discrirninator walk, The other one is the finite amount of a
charge needed to trigger the discrirninator, indicated by the shaded area .s
in Fig . 3 .8.
Qnantitatively the time walk e,ffect is represented by 116 - :
1
fit,
tonst . +
(3 .1)
vfii
where q is the amplitude of a pulse, and is a coeflicient determined by
the data . Thus tlw walk corrected time t is expressec by:
t = const . + t
v
(3 .2)
. /7r
where t is a measured time.
The walk is clearly seen in a scatter Plot ; (Fig . 3 .9a), which shows a
relation between arid Tiere t ' stamis for the time of Ilight between the
stop arid the tested connters, whereas q represents the integrated charge
of the signal fror the tested detector . As c,xpected, for pulses (fro' the
tested detector) wich higher arnplitude the discriminator generates the
Iogic signal cmfier, hence the rneasured time of ffighl, is Tonger .
CHAPTER 3. METIIODS OF TdfE DETECTO .1 E JAN A T I O N
3
0 .13
0 .11
U
0 .09
-~ 0 .07
29
exarnpfe of the scatter Plot t vs . * i eas red at the center
of the detector, A) before, B) after time walk correction
Figure 3 .9 An
For lang scintillation counters the shape and the amplitude of the pulses
depend on the hit position . Therefore the parameter /3 in equation 3 .1 anay
reveal position dependence . Table 4 shows this coe~cient evaluated for a
few points on the scintillator .
PM LI LG
c
SCINTILLATOP
PM P
Table 4.
/3 value [ns . pC 112
Pl\1 R
PI /l L
ßL
ßn
-16 .1
-10.6
1 6 .4
- 13 .0
-20 .5
-15 .2
-14 .2
-19 .5
position
A
B
C
D
As no regularity can be found in 0 variations, it is assm ed to be
constant over the entire scintillator surface . In further considerations is
taken as a mean value of the numbers in table 4, with the error being
a standard deviation of the mean . Evahlated /3 vali€es amonnt to:
~~ = - 18 .5 ± 1 .8 n.s•pC 112 and /3L -- -14 .0 + 1 .3 ns•pC' /2 . The inaccuracy
38
C1TAPTVR. 3 . A.lUT1101)S 01 , "1"UP: 1)VTI';C'1'«)1?. UXA IMNATION
in determining this coefficient implies timt the error of the walk correction
(o-At = o-fl 1 /Vij) is as high as 200 ps for Imv signals (sen Fig . 3 .9).
3 .5 Part ide identification
Firnis and protons, used in the experiment lind the sattle momentum of
2 GcV/c . However, Imcause of different masses, they h~acl different velocities givcn by :
pe
p c2
v =
(3 .3)
1p 2 + nt7,
rohere c is the light vdocity and rrzo,p are the rast; mass and rnomentum of
a particle, respe.ctively. This relation 1-eshits in the time of flight difference,
of the particles boing considered as 341 ps per nieten Thus, the dista,nce
of 7 m between the stop and the tested detector is eqnivahmt to 2387 ps.
200
250
150-
200
in
150
Z
0
0 100
PROTONS
50-
50
A)
0 22
23
24
25
26
TIME OF FLIGHT (ns)
27
Figure 3 .10 Time of Jlight spectnirn
8)
2 0 400 600 800 1000
ADC CHANNEL
An example of ADC spectrurn
for the tested deteetor
Figure 3 .10A shows a walk corrected time of flight spectrum between
these (7o-unters . As expected, very well separated peaks are observable.
To complete the considerations concerning the (hü), evaluation, a . typical ADC spedrum is showri in Fig . 310B . 'inke in the TDC spectra,
leere protons and pions are indistingnishable . This is becalise their energy
1oss is nearly the sarne .
Chapter
The response of the dete ion
module
The coordinate systern referrd to in the next sections is defined as:
AT
P
0,0)
Figure 4 .1 Definition of the coordinate systern
4 .1 The detector efficiency
One of the main detector r :haracteristics is its vfliciency . This is defined
as that fraction of events irnpinging ein the detector, which is registered,
An efficiency is a firnction of radiation type, cnergy and of the detector
material . This is becanse no material is eqirally sensitive to all radiation kiruls, and the sensitiveness of (\ach material (langes with the given
radiation energy,
order to be registered a particle has to pro(we a signal, which is high
39
40
CHAPTER 4 . THE RESPONSE OF THE DETECTION MODULE
enough to exceed a necessary discriminator threshold . The signal height
is associated with the nurnber of photons generated in the scintillator,
which is proportional to the energy deposited by the particle . Thus the
efficiency can be expressed as the probability of producing enough photons
to generate a Signal higher than the threshold.
For a monoenergetic beam the variations in the number of produced
photons is governed by the Poisson distribntion . Consequently the efficiency can be approximated by :
,-=oo /l t
(4 .1)
,-ith
r.
where th is the number of photons equivalent to the 'threshold pulse' and
u is the average number of produced photons.
Minimum ionizing particles deposit about 0 .7 MeV in a 4 mm thick scintillator . The average energy required to produce a scintillation photon is a
fixed number dependent only on the scintillator material . In plastic scintillators it amounts to about 100 eV per produced photon [17j . Assurning
(in the worst case, see section 4 .3) the light collection efficiency 7%, and
the photocathode efficiency 25% one can calculate that on the average 122
photoelectrons are generated.
7 - 105eV
7 103 sezntzllatzon
photons
-4
t.,d
122 photoelectrons
At normal parameters used for the photomultiplier one photoelectron gives
a signal with a height in the order of 40 rnV . Therefore, a signal caused
by 5 photoelectrons easily exeeds the discriminator threshold of typically
50 mV . In the case being under consideration 5 photoelectrons are equivalent to 286 scintillation photons . Taking in equation 4 .1 th ,- 286
and a = 7000 one obtains the efficiency better than 99 .99% . Moreover,
the efficiency of the scintillation counters for mirÜmum ionizing particies
detection remains constant down to a scintillator thickness of , 0 .3 mm.
Then fit is equal to h, 525 .
CIIAPTE1?. 4. 7'IW 1? .1„SPONSP; 01 , Till°,' 1)1';T1 C'1'ION
4 .2 A pulse charge variation
Althongh in ternary scintillators seif absorption is relatively stnan, für
long detectors it influences significantly the light collection process . Therefore, für scintillators of clirnensions cornparahle to the hink )igitt attennation 1ength orte observes a variation of the pulse lteigitt depending on the
irradiated position, (wen für mormenergetie partielos.
I"n hing scintillators the photons normally make a lot of reflections before they reach the (Age coupled to the light guide . Thus, the balle light
attenuation is not fite only factor causing ehanges of the signal height.
The other orte is the reflection lass at a scintHlator surfilces.
These factors result in a Iigla attenuation deserihed by :
N = No
e3_ ,
(4 .2)
where
is called a technical light anennation Iength, and 1V0 , N are
the mmiber of collected photons when the scintillator is irradiated at
X = () and x, respectively. Figure 4 .2 shows the relative photornultiplier
100
-G
-4'
..---,. 90
+
u.J
o+
--FM L
+ o
PM R
c~j
0+
+OOQ_
▪
+
-80
.pQ0
40t ++
c) 70
L,J
+
cn<
[1_ r2L
0
z
o+
o
+
+
++-+-++ , °
Oq-
P0i-
60
o
pions
(4;1
0
cp
+ o
++
+
+
50
0
100
eiectrons
200
300
X (mm)
400
Figure 4 .2
Relative variations in the signal charge
für 2 GeVie pions arid für electrons from a "Sr source
12
CIIAI > TVI?. 4 .
1ZESI«)NS1'',
TIM
AIODULV
pulse charge variations according to the position along the Ihre connecting
the photorrmItipliers . The signal charge was evahiated as a trican value of
a Gaussian distribntion fitted to each ADC spectrum . By fitting a line
to the experimental points (shown in Fig . 4 .2), wiChout Caking the data
at the ends of the scintillator into account, the technical light attermation
length was fonnd to Ire eiqual to 88 cm 15 cm.
400
500
ui
400
z
z
z
300
1
0
0 300
C,J
4
y
50mrn
y 5mm
200
0
300
X (mm)
Figure 4 .3 A rough
position deterrnination
4-5o
2000
150
300
x . mm)
'e
e.
450
Figure 4 .4 Pulse charge dependence on the y-coordinate
The considerable variation of the charge frorn the pulse irr the x-direction allows to estimate roughly the x-coordinate of the hit position.
Figure 4 .3 Shows the signal charge expressed in ADC«°hamtels, together
with the estirnate(1 error of a single measurement (o of a Ganssian distribution) . Fror this figure olle can derive that the error of determining the
hit position amomas to + 13 cm . Taking into accomtt, however, that irr
the actual experirnent the scintillator will he read out sirmtltaneously by
two photorrmItipliers this value will reduce to + 9 cm.
In contrast to the pulse height variation in the x-direction, there is
essentially no dependence on the y-coordinate (sec Fig . 4 .4).
Sirnilar tests of pulse charge changes with respect to the hit position
were performed with two other scintillation deteetors, which are shown
CIM PTIM 4 . TUE RI ,.:SI«)NSP 01;"l"lIr 1)WIT:CTION MODULE.
43
schematically in Fig. 4 .5 . As can Im seen, a scintillator with read out on
both ends reveals, in one photornulCipli( , r, nnieh ',arger pulse height variations in comparison with those read out on one edge only . This is because
in the Inaer the light refiected front the sicinfinator edge (c)pposite to the
end gIned to the light quide) also contributes to the signal arnplitude.
o0-fg-o ts.,
+
0 000 0 0
+
90
A)
LG
A
0
+
sc IN.
B
+
B)
PM
4-
----''---
LG SCINTILLATOR
. .. .
Ft
Figure 4 .5
250mm
50
150
»i0O
X (mm)
++
+
,
450
Comparison of pulse charge changes for varions detectors
4 .3 Estimation of the light collection efficiency
The considerations in this section concern the measurements performed
with a 9'Sr eleetron source.
Dividing the integratecl Charge of the signal by the photornultiplier gain
one can calculate the nutnber of photoelectrons generated in the photocathode (sec Fig . 4 .7), which allows to estimate die numher of photons
reaching the photomnultiplier . On the other hand one (-an estima .te the
nurnber of photons produced by the registered partiale in the scintillaton
Electrons when passing through matter suffer a collisional arid radiational cnergy loss . flowever, at energies of a few MeV or less the latter can
negleded . Figure 4 .6A shows tim ionization energy loss per nnit length,
for electrons in a scintillator material, derived front the Bothe Bloch' for-
i coeiTicients for plasfic scinCiliators can l~e found in 13] on Page 2$
IMSPONSE
44
1)11 1- VG770N MOPULK
nmla . This relation allows to calenlato (numerirally) an average energy
loss frorn eleetrons within the scintillator tuodium of Ute testrd detector.
> 1,1
1
r
r
P
S€ r r
_r_L L
e k£
f
-r..-1-
Irr -
r
(V
.
-I-L
t
,._.,
f
6
r
r
1
X
1
1
r
L_--€-__L _t ._l3 _ r
1
1
r
r
6
3
r
r
1
+f-
•r..l
t
eir e
1£
L
1£ I s
r
6
1
6
0
t! 1
r e
1
_ -1-
1
c 1 .0
Lt
z
iil
1
0 .9
r
(7i
r
r
1
0
EL
w
e
0
1
r
1
ELECTRON ENERGY (MeV)
0 .8
L6 1 .7 1 .8 1S 2 .0 2 .1 2 .2 2 . .3
ELECTRON ENERCY (MeV)
10
Figure 4 .6 A. The stopping
Power dE/dx as a function
of energy for electrons [18]
B. Average energy deposited by
electrons within a 4mm thick
scintillator versus its energy
By a suitable threshold adjustrrtent, frorn the coMinuons (3 spectrurn,
only electrons of an energy Aue 2 MeV were registered . Due to the Iow
energy resolution of the scintillator the energy of the measured electrons
is not very well define(l . iowever, the energy deposited by eleetrons of an
15
350
o
z
1
o
0w
-
++
300
PM R
+ --FM L
13 0
+
w
+#
+
+
° 250
m
n
+
11
+
+
0
z
#
++
o
+
200
w
++
+
▪
+ +
150
300
m
--n
L,-
+
y 50mm
'1500
9
++
f-
+,
+
m
z
+
-
7
n
--<
450
X (mm)
Figure
4 ..7
Number of electrons an the ferst dynode and Iight collection
efllciency as a function of the position in a scintillator
(]!!A fli7
d\.
j?! SPO SP O!
' 1 ' 111 '
2
!»J
1 t P 7 " 10 MODULE
45
energy larger than 1 .0 Me r is nearly ccinstaarrt and aa .rncaunts to r0 .93 Med.
(see Fig . 4 .GB) . Thris, on the average one electron preaelrices 9300 fl .uorescence photons . Hence, clivicling the n.iimber of f>hcatoelectrons by 9300
and hy the pl tcacaa,theacle eiciency (0 .25) one ohta .ins the light collection
efficieiicy (sec Fig . 4 .7) . For one plLEatornriltiplier this varies, depending
on the hit position, fror r7 % to 13 f0 . Consequently, the fraction of
flnorescence light seien hy hoth photomultipliers amounts to ' 20% incle®
penclently of the hit position in the scintillator.
4®4 The light
lse velocity
28
30
Figure 4 .8 'Firne of Rigid spectra between the stop and the tested counter,
for two diWerent beain positkns, measalred with 2 GeV protons and pions
Figure 4 .8 shows thaat the time of the sigitaa .l arrival at a TDC rinit
clepemIs on Hie x-coorcliaoitc cif the irnpaact position . Thus, in order to
esta.hlish the time wheu the partiele strikes the cletector, it is necessary
to kirow the hit position as well as the ef ective velocity of the light pulse
in the seintillaator . In experiments at the COSY ring the for°rner will be
providecl hy the dri .ft ehambers (see Fig . 0 .1).
The mein va .lue of the Gaussian distribution fitted to each speetrum
is calculated with Qi very sm all error iii the order of 7 1 )s, which is defined
46
CITAPTER TUE RESN.»IST; OF TUE DETECTION MODULE
Vv/,
where v is the variance of a distribni,ion and n is a number of
events . Unfortimately, the TDC peaks are shifted becanse of the time walk
in the discriminator . This shift varies over the scintillator surface, since the
average signal height depends an the position . Thus, in order to rneasure
the real time differences between the TDC spectra originated in different
positions, the time walk has to he corrccted . Inaccuracy of Ulis correction
as
estirnated in chapter 3,4 amonnCs to values as high as 200 Iss ( o ) for Iow
Signals.
3 .5
3 .5
•
L ELECTRONS
3 .0
2 .5
*
BEFORE OORREC1ION
äe
AFTER CORRECflON
P L4 R
3 .0
*
*
P L
o
2.
all
*
c2 .0
C 2.
** md
*
. gw
*,.
1
.5
*Bi
0
0 .5
0.
Y
= 50 m m
Figure
0
450
X (mm)
4 .9
•
0 PIONS
5g
n.
tl
A)
ao
B)
150
erleo
ELECTRONS
300
X ( M IM )
n ie
450
Variation of the TDC peak with respect to the hit position.
A.
For electrons before and alter time walk correction,
B.
For electrons and pions after correction
Time is measured in relation to the timing of the signal produced at x 25 mm for PM L and at x = 425 mm for PM R
By fitting a 1ine to the mea .snred points the Iiglit signal velocity was
found to be:
'Fable 5.
Light )nlse velocity cm ns
15 .5 .+ 0 .6
15.3 + 0 .6
- 0 .8
easurements with:
90 Sr
P
R
Pl\
source
PM R.
pions and
12 .3 ± 0 .6
PM L
protons
CHAPTER 4 . THE RESPONSE OF TIIE DETECTION MODULE
47
The first three values in Table 5 are equal within the error lirnits,
whereas the velocity measured by means of PM L at p and 7r + beam differs
from the rest . This discrepancy might have been caused by a wrong TDC
calibration or by pedestal variations during the measurements . Since the
experimental set-up is dismounted further inspection is not possible.
T . Tanimori et al . [161 exarnined a similar scintillation detector . They
obtained the effective light velocity of 16 .3 + 0 .2 cm/ns for 2 cm thick seintillator and 16 .7 + 0 .2 cm/ns for the 3 cm thick one.
4 .5 Time resolution
Fluctuations of the signal timing are caused mostly by a walk and a jitter . The former can be corrected by the data analysis, conversely, the jitter
arising from the intrinsic detection processes can not be avoided . Because
of this even two exactly coincident signals do not give the sarne time . More
over, the accuracy of currently available TDC units is not bettet than . 25 ps
per channel . Thus, to take advantage of the signal timing, the System (detector + electronics) time resolution has to be estimated . This was dope
by measuring the distribution of the time differences between signals from
PM-L and PM-R, generated by the saure particle . Ideally it should be a
b-function, but obviously it is not the rase, in fast its width changes with
irradiated position . A typical distribution for At =2: tR - tL , together with
the fitted Gaussian function, is shown in Fig . 4 .10 . The variance of this
distribution resulting from fluctuations of the signal timing from the two
photomultiplies is expressed by :
(4 .3)
where, and are the standard deviations in determining tR and
tL, respectively . In order to establish the overall time resolution of the
tested detector this variance has to be cornbined with the error of a walk
correetion .
48
CI'IA PTT i?.
0
2000
TFM RES1«)NSE 01"
3000
A t (ps)
Figure 4.10
4000
1)Ul'iX7 ION MOI)IJI,E
5000
Art exnimpfe of a
spectrum
measured at x = 25 mm
The rdntimn between the measnred hit time and the admal one is as
folloWs :
x
pj(
) + cont,.
Ls
t rrira .eltred t
Vr f f
ll
mra .qured
x ) + const .
walk «jVeff
(4 .4)
where, x is the hit Position, and L is the scintillator length . The " start"
moment for the time of ffight mcasnrement between the S1 and S3 detectors (sec., Fig . 0 .1) in the COSY - 11 (E-5) experiment will he defined as
a mean valne of 1; 1, and tn.
t u.
tn
ttart
22?.ca.5ured
nteasure
''--+cortst .
(4 .5)
2
Walk
This irnplies that
(7 2
i t(aT
t7
'
(7i2
rt
(7
Lit'welk
L
(7 2
Al n
2
1;
EJy 81 /1
PIER (L
TU1 DE7TCTION AIODIJM
.
19
Term cr t2L -1- 12n is eqiial to the varianee of the At=t 1,-t ll distribution,
which reveals a . dependence on the irradiated position . Similarly the standard deviation of a time walk correction (o- A1,f,,k ) is a function of a signal
height (see seetion 3 .4), heitre depemIs on the hit position . Consequently,
the evahiatcd vahte of o- t,t„ (shown in Fig . 4 .11) is not constant over the
scintillator surface .
.
300
(n . 250 :2
ca.
+
uJ
i 5Q
300
450
X (mm)
Figure 4 .11 The overall time resolution of the tested detector for a few
points in the scintillator . The curve is drawn to quirle the eye
This estimation Shows that the error of determinig t, iarf varies from
160 ps at the scintillator edges to 220 ps at the centre . Furthermore,
for the resolution of a time of ffight set-up, this approximated has
to be folded with the time resolution of the stop rohliter .
ACKNOWLEDGEMENTS
51
Acknowledgement s
1 would like to thank to Prof. L . Jarczyk and Prof A . Strzalkowski for
supporting my work, and also to Prof. K . Killan and Dr. hab . W . Oelert
for enabling me study in the Nuclear Institute of the Forschungszentrum
Jülich . 1 am especially grateful to Dr . hab . W. Oelert for guiding me
during my work, and for making my participation in rneasurements at
CERN possible.
1 owe a great debt to Dr . D . Grzonka and Dr . T . Sefzick for introducing
me into the technique of measurements, and for their kind answers to my
continual questions.
1 wich to record my thanks to J . Thimmel and M . Wolke for many
stimulating discussions, and also to A . Misiak and J . Majewski for their
help in dealing wich electronics apparatus.
Dr . M . Ziolkowski and K . Pysz were very kind in showing me how to
use some programs necessary in rny work.
1 would like to thank the many people who haue helped me in many
other ways . In particular 1 am indebted to : Dr. R. Bröders, M . Dahmen,
K . Dickinson, Prof. Ben Gibson, M . Kistryn, Dr . S . Kistryn, M . Rook,
Dr . E. Roderburg, R . Stratmann, Dr. A . Szczurek and P. Zolnierczuk.
Finally rny very special thanks are due to Dr . hab . W . Oelert and
Dr. D . Grzonka for having read the text of the work and for their helpful
comments and correetions .
52
REPERENCES
References
[ 1]
W . ®ehrt, Threshold tl eas arerraents at the Experirnent CCSY - 11,
Talk presented at the 105. WE - Heraeus -Seminar,
KFA IKP(I) - 1993 - 1.
[2]
Kay Königsmann, New I 'adronic states,
CERN-PPE/91-160, 1991.
[3]
BICRON,
Prernirarra Plastic and Liquid Scintillators,
BICRON, Plastic Scintillators a Srarvey,
by C .R. Hurlbut, 1985.
[4]
J . B. Birks, The Theory and Practice of Scintillation Counting,
Pergamon Press, London, 1964.
[5]
B . Bengston, M . Moszynski, Study of Primary Energy Transfer Process in Ultrafast
Plastic Scintillators,
Nucl. Instr. a.nd Meth . 155 (1978) 221-231.
[6]
M. Moszynski, B . Bengston, Light Pulse Shapes vorn Plastic Scintinators,
Nucl . Instr. and Meth . 142 (1977) 417-434.
[7]
M. Moszynski, B . Bengston, Status of Timing wich Plastic Scintillation Detectors,
Nucl. Instr. and Meth . 158 (1979) 1-31.
[8]
Röhre GmbH, Datenblatt Plexiglas,
Darmstadt, 1985.
[9]
G . Keil, Design Principles of Fluorescence Radiation Converters,
Nucl. Instr . arid Meth . 87 (1970) 1.11-123.
[10]
[11]
L . Thornhill Scintillation Counter Technology,
Geneva, 1983, SB/AC/ST/3173.
THORN EMI,
[12] Glenn F . Knoll,
Photomultipliers, 1986
Radiation Detectiora and Measurement.
R.Ei''1ZENCiS
53
[13] Williarn R . Leo, 7eclanique,s for Nnc..lear• and 1'rarliclr l'hysirs l ::rperimenls,
Springer - Verlag Berlin Heidelberg, 1987.
[144
'1`. Selkick el, a.1 ., ii srlslern fror Simulation of Scinlillettor l.ighl Signals,
Nucl . Instr . and Meth. in Phy. Re. A288 (1990) 571-573.
[15]
A . Marin, G. Visrnara, The llelatl Wirr Ghettoher 1lescripiiori,
CER .N - 1,KP Division 1,Ki'/131-TA /Note 85-3.
[16]
Tan i rnori et, aal ., A Test of 150 r.rta x 20 rm Wide '!`irr:e-of- Plighl Sein!illatiou
Counters,
Nord . Instr. and Meth . 216 (1983) 57-65.
[17]
Parf,iclc I)atn . Group, ierrierrr of 1'arlicle !'roperties,
Physica.l Review 1)45, ParL2, 1992.
[18]
A . Ernpl, Untersuchung eines Szintillaiion - Detektors besonderer Form,
Diplomarbeit, Bonn, 1989.
Furt her Reading
A . Bndzanowski et a l, GOSY - 11 Deicr.tor Arrangement
'U -1KP, 111-25911, 41.
L. .Iarczyk et a.1., lieconsiruelion of ihe Shape of i feson liesonanees Produrcd in
ihe pp ---> pp/1f Reaciion Neoar 7 'hr-eshold,
KFA-I K P, Jül-2590, 43.
M . Moszynski, Studrl of l,ighi Colkction Frrnn C7i/indrr,cal Scinlillalors,
Nucl . Instr. and Mel,b . 734 (1976) 77-85.
3. B. Birks, Fncrgy T •ansfer in C)r ;ganic System,
3. Phys . 13 (l'roc. Phys . Saat .), 1968, Scr . 2, vol . 1.
E. C_;atti, V. Svelto, Revier of Theories and Experiments of la'esolniraq 7'irrae wich
Seiniillation Counters,
Nue1 . Instr . and Mette . 43 (1966) 218-268
T. Sugi al,c et, aat, 100rm lang Time-of-Flighi Scintillation Counters witla RATS
Resolution of 50 ps,
Nucl. Instr . and Meth . in Phy . Re . A219 (1986) 354-360.
J ECR0V, Research Instrramenlatioru C,`ealalog, 1992 .
.' ..
RSCHUNGSZENTRUMJULICHm# H
:
: :
...::::..- : .:.:-.. :-.:, . ::-:. :
:
Jül-2825
CCober 1993
1SSN 0944-2952