My Abstract

(1)

IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 38, NO. 2, APRIL 1991

The

BaF2

Photon Spectrometer TAPS

R.

Novot ny

1I.Physikalisches Institut Universitat Gieflen for the

TAPS-Collaboration (GANIL-GIESSEN-GSI-KVI-MUNCHEN-MUNSTER)

Abstract

T h e detector system TAPS ( Two/Three A r m Photon Spectrometer ) has been designed and installed t o study high energy photons as well neutral mesons produced in relativistic heavy ion reactions. The spectrometer will consist of u p t o 384 individual BaFz-modules packed in arrays of 64 scintillators. Each spectrometer arm, which carries two detector blocks, can be moved around the target location independently. The design concept, tlie specifications of the individual scintillator and test results of sub-arrays performed with monochromatic photons and charged particles will be presented. First experiments a t GANIL an d SIS (GS1,Darmstadt) have been performed exploiting heavy ion beams up to a projectile energy of lGeV/ u.

1 Introduction

T h e photon spectrometer TAPS (Two Arm Photon Spec- trometer) [l] has been designed t o detect low and high energy single photons as well neutral mesons such as T O

and ipmesons. These probes can be exploited t o study reaction dynamics in relativistic heavy ion collisions (for example a t the SIS/ESR facility a t GSI, Darmstadt) or the excitation of sub-nucleonic degrees of freedom in photon induced reactions, using the new tagged photon facility a t the electron accelerator MAMI B in Mainz.

Since neutral mesons decay predominantly into two photons ^{( K O :}99 %; 11: 39 % of the total decay) the invariant mass reconstruction requires the coincident detection of both decay photons. The relevant momentum distribu- tions of mesons produced in medium energy heavy ion reactions will lead t o high energy decay-photons up to E, ^E1 GeV emitted a t relative opening angles between the photon pair of up to 0 1 2 5 180'.

As a consequence, the TAPS-spectrometer concept coni- prises u p t o three movable towers which can be ro- tated around the target position independently (angular range with respect t o the beam axis: 10'5 @ l a b 5 1 6 5 " ) a t a variable distance to tlie target ranging between 0.5 ni

5

D

5

3.5 m , respectively. T h a t allows to cover in several detector settings all required relative angles

Fig. 1 . Schematic view of two arms of the photon spectrometer T A P S .

between two coincident photons in order t o reconstruct the full momentum distribution of the detected mesons.

Each tower carries two detector arrays consisting of 64 scintillator-modules. The relative angle between these blocks and their distance to the target can be varied in- dividually. At a distance of 1 in to the target each array (front face: x.y z 51.42 c m 2 ) subtends an angle range of AQz 29', and

M y

z 24'. respectitcly. T he ceiitial module of each block can be positioned a t relative angles between 27' and 120'. All position adjustments can be read out digitally and will be set remote controlled in a later stage. Fig.1 shows schematicallv the set-up of two spectrometer arms.

Clean photon identification requires an efficient discrimination against charged particles.That can be achieved via the-of-flight measurement, pulse-shape analysis of the detector response and a charged particle veto detector of the same granularity mounted in front of the BaF2- crystals.This veto-system provides a n additional on-line particle suppression.

0018-9499/91/0400-0379$01.00 0 1991 I E E

(2)

T h e invariant mass of the neutral mesons is determined by

T?Z,,,crmC2 = d 2 E 1 E 2 ( 1 - COS012) (1) E l and E2 refer t o the energies of the observed photon pair subtending a n opening angle 0 1 2 in the Lab system.

Complete conversion and reconstruction of the electromagnetic shower developed by the high energy photons deter- mines the achievable mass resolutions.The properties of the detector material as well the shower leakage o u t of the detector volume limit the energy resolution. T h e spatial information is primarily given by the module size which has to accept high particle multiplicity in heavy ion reactions a s well. T h e determination of the point of impact can further be improved by evaluating the center of gravity of the electromagnetic shower. Based on these require- ments, BaF2 has been chosen as the only adequate scin- tillator material exploiting its excellent properties which have been investigated in various applications [2,3,4] and will be documented in the test measurements described in the following.

2 The detector modules

T h e geometry of the BaF2-crystal and the fully assembled individual detector, respectively, are shown in Fig.2. Each module consists of a hexagonal crystal'

crystal geometry

dlmanrlon- In mm

full rss-mbly

provided by PVC ihrlnklng tub.

Iangth L

-

^{2 5 0}

of aluminum foil (15 p m ) serve a s reflector. For time and energy calibration and stability control light of a N2-1a.ser3 (wavelength A = 337nm) is fed into each detector by a quartz fibre glued into a groove machined in the cylindrical end cap of the BaF2-crystal [5].

The photomultiplier tube4 is coupled t o the crystal with silicon grease5. The whole assembly including the photomultiplier base housing is contained in a light-tight shrink- ing tube which in addition gives sufficient mechanical stability. A cylindrical magnetic shield' surrounds completely the tube and the cylindrical section of the crystal a n d provides effective magnetic shielding up t o a flux of 0.02 T .

PMT Valvo xP2972

E

NE102A scintillator

Fig. 3 . Schematic viw of an individual CPV-element ( a ) and the full assembled array in front of a BaF2-block.

quartz fibre for W - l a s e r callbratlon and stablllsatlon

.

\

^magnetic^shleldlng

1.5mm Permenorm S4

The assembled detector

Fig 2 .

detector module.

Geometry of a BaF2-crystal and a fully assembled

(length = 25 cm = 12 XO, XO: radiation length) with a cylindrical end part

(4

= 54 m m ) . All surfaces are opti- cally polished. Eight layers of 38 p m P T F E 2 and one layer

~ _ _

Firma Dr. Karl Korth, Kiel, FRG

*

Tetratec Corporation, Feasterville,PA, USA

A charged particle veto-detector (CPV) [6] ^ofidentical granularity, split into two sub-arrays of 32 individual plastic detectors, is mounted in front of each BaF2-block.The hexagonal scintillator (NE102A, width: 651nin (diameter of inscribed circle), thickness: 5mm ) is read out via per- spex lightguides (length: 19-53 cni, width: 27mm ) of the same thickness coupled to a photomultiplier7.The slightly overlapping scintillator elements are arranged in three lev- els. The maximum plastic thickness in front of a n individual BaF7-modnle can add up t o 24 min (0.05 Xo) including lightguides and reflector material. Fig. 3 illustrates the assembly of an individual module and the arrangement in front of the BaFz-array.

3VSL-337, LSI Laser Scientific, Inc.

4Hamamatsu R2059-01 5Baysilon 300.000

6Permenorm 5000 S4, VAC Hanau, FRG 'Valvo, XP2972

(3)

performance parameter (mean value)

3 The BaF2-detector

accepted specif.

crystals limits (310 samples)

T h e determination of energy and time response of single detector modules or sub-arrays has been performed using standard nuclear electronics. In case of time-of- flight measurements constant fractzon dzscrzmznators and tzme-to-dzgatal- or tzme-to-amplatude-converters have been used, depending on the required resolution. T he energy information is deduced by integrating the light output of t he scintillator in a charge sensztzve A D C (integration width w =2 p s (total light output),w = 40ns (fast component)). T h e final TAPS electronics system uses several highly packed C'AMAC modules [8] which have been developed recently within tlie TAPS collaboration.

I A E j a r t l E

[%I

~

A E I E

l%l

3.1 The crystal specifications and accep- tance tests

T h e performance tests [9] of each crystal involve monitoring the

0 UV-transmission a t various points perpendicular to the axis of the crystal in the wavelength range between 185 and 500 nm using a double beam UV- spectrometer' and the measurement of tlie

0 energy response and pulse shape determined with low energy 7-sources.

T h e required optical performance of the crystal is speci- fied by limits set on the transparency a t the wavelength of the major scintillation components ( A % 200,220,300 nm).

34.6 45.0

11.5 12.5

I % , 6

' fast/slow a t At=40ns

1

^9.5 7.0 Table I . Obtained experimental test results of 310 accepted c r y s t a l s in comparison t o t h e required spec i fi e at i o n s

Table 1 shows the obtained transverse UV-transmission (crystal width 59nim) corrected for Fresnel reflections and

the deduced attenuation length A ( Itranrm=Iinc ).

T he results are averaged over 310 TAPS-crystals fully accepted by October l s t , 1990 [7] Those crystals, which show distinct absorption bands due t o impurities (like P b2+ [lo]) causing a deterioration of the luminescence properties

,

have been rejected.

900 n

;

⁸⁰⁰

td

700

600

Q,

E

⁵⁰⁰

E

400

n 0

U v)

+ 300

ld 200 y.

100

p h o t o n s

-I

total light o u t p u t ( a . u . 1

Fig. 4 Pulse shape analysis of BaF2 signals due t o intrinsic radioactivity.

The inspection of the pulse shape is quantified by the amplitude ratio of the fast coniponent t o the slow scintillation response measured a t a delayed time ( At=20ns, or 401is, respectively)

.

A qualitative measure of the signal shape can be deduced froin tlie sensitivity of

CY

separation (intrinsic radioactivity due to contamination [ I l]) based on tlie pulse shape analysis (see Fig. 4).The absolute amount of light output of the fast and slow component, respectively, is expressed by the obtained energy resolutions (F WHM) for low energy y-sources. T h e corre- sponding average values are listed in Table1 in comparison t o the specification limits.

3 . 2 Response to electrons

T he response t o electrons [13] in the energy range between 10 and 13 5 hIeV has been investigated using the electron linear accelerator at Giefleii [13]. Since electrons above tlie critical riieigi of E, 2 13 RIeI' in BaF2 create electio- iiiagiietic showers t h e central liexagonal crvstal has been surrouided bv six identical detectors. A n active collimator consisting of' thin plastic scintillators has been used to define the point of impact and to serve as a time reference.

5 shows the line shape of tlir central detectoi, tlie surrounding ring and of the entire array a t

As a iesult, Fig.

'Hitachi U3200

I

(4)

25

c 20

- ²

¹⁵

9

¹⁰

8

4

5 15

h

$

g 5

v 10

e 01 c

-

25 ²⁰

- ^z

¹⁵

9

¹⁰

8

e

5 0

I

<E> = 33.6 MeV

neighbouring detectors

<E> = 5.4 MeV

2 0 0 - U1 Y c

100

I

Eped = 182 MeV

I

AE/E = 6 % -

L .

1

^IO

^j

( l i . 2 f 1 . 3 ) , 22

I

(14.lf0.7)

Fig. 5 Line shape of the central detector ( a ) , the ring ( b ) and the entire array ( c ) f o r 43.5 MeV electrons. The dashed line represents the results of GEANT3 simulations.

(12.8f1.5) -

(12.9f0.7)

43.5 MeV electron energy in comparison t o GEANT3 [14]

simulations. T h e influence on energy resolution of discriminator threshold, energy leakage and dead volume due t o detector housing has been studied over the entire energy range. Table2 summarizes the deduced results in comparison t o the simulations which d o not fully account for energy losses due t o a n imperfect detector geometry.

1

Beam

I

Energy

1

Time

1 ’

32 43.5

1 I

Resolutfon [%]

I

Experiment Simulation

i12.3f0.6j i11.7f0.7j 220f20 (12.9f0.6) (10.7f0.6) 160f20 Resolution’

4 5 0 f 2 0 1 7 0 f 2 0

[PSI

j

Table 2 . Experimentally measured energy resolutions ( F W H M I E ) for the seven module array in comparison to GEANT3 simu1ations.l The time resolution (FWHM) contains the intrinsic resolution of the plastic start counter (Ate1 Oops).

T h e time resolutions (FWHM) deduced froin time-of- flight measurements of the central BaFz-detector relative t o t h e plastic start counter are below A t z 200 ps at the highest energies (see Table2). An optimum value of A t = 110 ps has been obtained in a n improved final con-

figuration at the same electron energy.The FWHM of the neighbouring detectors increases t o 800 - 1000 ps due t o the low deposited energy ( 3 - 5 MeV) and the varying time and location of secondary shower components scattered into these modules.

2 0 . * v J - L u f ,

J

R

s u m of 7 detectors

5 0 .

s u m of 19 detectors

250

0

energy [ MeV ]

Fig. 6

of the central detector, 7 and 19 modules, respectively.

Energy response to tagged photons of E7 = 186 MeV

The beain spot of zz 1 cin in diameter was defined by the active collimator. The reconstruction of the impact point has been obtained from the center of gravity of the energies deposited into the individual modules by taking into account the shower geometry. The position resolution aiiiounts typically to Ax,y zz 3 cni ( F W H M ) without correcting for the beam profile.

3.3

In a test experiment a t the Mainz LINAC a colli- mated beam

( 4

^z1.2 cm) of tagged photons of energy E, = 186 MeV has been used t o determine the energy response of a TAPS sub-array comprising 19 BaFz-modules

Response

to

high energy photons

~ 5 1 .

(5)

c ”’

^p ^d ^t

1

CI

L -

O m -

I I I I I I I I 1

wide gate 6aFa C a . u . 1

Fig. 7 Scatter plot of the fast scintillation component versus the total light output. Photons and charged particles are separated by pulse shape analysis.

Fig. 6 shows the line shape of monoenergetic photons (E, = 186 MeV) impinging on the central detector module.

T h e achieved energy resolution of A E / E = 16% (FWHM) can be drastically improved by adding in the shower leakage deposited into the first (6 modules) and second (12 modules) surrounding rings of BaF2-detectors, respec- tively.The resulting energy resolution of the array reaches AE/E = 6% (FWHM) which is in agreement with expec- tations based on GEANTJ simulations. Contributions to the spectrum due t o pile-up and electrons from conversion in t h e Pb-collimator have not been fully rejected. T he mean energies deposited into the first and second ring of surrounding detectors amount t o E,,,, = 19.0 MeV and 5.6 MeV, respectively. Similar t o the results with monoenergetic electrons spatial resolutions of Ax,y x 2.5 cm have been obtained as preliminary results. T he size of the beam spot has not been unfolded. However, in order to repro- duce the absolute point of impact a more refined procedure has to be developed to take the hexagonal crystal shape and the electromagnetic shower profile fully into account.

3.4 Response t o charged particles and neutrons

TAPS-detectors have been exposed to light charged particles produced in p- and a-induced reactions a t the KVI cyclotron in Groningeii [lG].Targets of C,ZnS and Pb have been bombarded by 55 MeV protons and 100 MeV CY-

particles. T h e BaF2-detector was mounted a t a scattering

IO00 900

000 4 m 700

600 500 400 30U 200 I O 0

d

i---r-r--i--r--n

I

hydrogen

\

helium

_j

I

gammas

/

200 400 GOO 000 1000 1200 1400 I G O O 1000 2000

particle identification [ a.u. ]

Fig. 8

analysis as given in Fig. 7.

Projection of the particle separation via pulse shape

angle @lab x 45’ outside the scattering chamber t o detect light reaction products passing a 25pm Kapton foil.

7 illustrates the photon a nd charged particle identification exploiting pulse shape analysis.The fast component and the total light output of the BaF2-signal 1i.ave been integrated separately in charge sensitive ADC’s. Well separated bands can be observed and identified as photons, Hydrogen and Helium isotopes, respectively. Fig. 8 shows the photon/particle separation after a projection along the bands marked in the previous scatter plot.

9 demonstrates the excellent energy resolution achieved for protons inelastically scattered on 12C mea- sured a t a projectile energy of E, = 55 MeV. T he deduced energy resolutions (FWHM) amount to A E / E x 2.4 % for 40 to 55 MeV protons, respectively. A comparable result of AE/E x 3.5 % has been determined in case of 82 MeV a-particles.

The identification and discriiiiiiiation of high energy neutrons is based on the complete infcrniat.ion p r o ~ i d e d by t.iine-of-flight. response of the veto-det,ector and 1)ulse shape analvsis of the BaF2-signal. Low energy neutrons (far below 100 M e V ) int,eract within the detector material predominantly via (n, ? )-reactions [IT] and can be dis- criminated only by a t,ime-of-flight ineasurement. Using t,he secondary neutron beam a t the SATTTRNE facility a t Saclay !he response of BaFzt,o neutrons in the energy range bet,ween 200 MeV and 850 LleV has been invest,igated ns- ing a sub-array consisting of seven 20 cni long prototype BaF2-detectors [18].

T he pulse shape analysis confirines tha t (n,p) reactions play the dominant role in the studied energy region. There-

The scatter plot in Fig.

Fig.

I

(6)

150

120

U v)

90

V

60

30

'10 15 20 25 30 35 40 45 50 55 60

energy [ MeV ]

Fig. 9

the reaction 12C(p,p')12C at a 55 MeV bombarding energy.

Energy response of a TAPS-detector to protons from

fore, neutrons which d o not trigger the CPV-system will be identified as protons applying pulse shape analysis.

T h e kinetic energy can be deduced from the measured time-of-flight. In addition, the deposited energy (photon- equivalent) decreases from

-

⁴⁰^%^{a t}^{E, =}200 MeV down t o FZ 24 % at E, = 850 MeV mostly due t o energy leakage.

An overall detection efficiency for neutrons (for a n energy threshold of 5 MeV) of E = 1 7 f 3 % can be deduced and stays nearly constant over the studied energy range.

4 First Experiments

T h e TAPS photon spectrometer has come into operation beginning 1990. After first experiments at GANIL, which have been performed in a spherical arrangement of 19- module detector blocks, two completely installed spectrometer arms have been used t o take first d a t a a t the new SIS-facility. Fig. 10 shows the experimental invari- a n t mass spectrum obtained in a first experiment a t GSI t o investigate the T O production probability in the system "Ne

+

27Al a t 350 MeV/u [19]. Fig. 11 documents the quality of photon/particle discrimination via pulse shape technique. in a n on-line spectrum of the reaction of l G e V / u 40Ar

+

,atCB. The separation appears t o be efficient u p t o the highest energies covering the range of the total light output up t o z 400 MeV of photon-equivalent energy. T h e calibration is based on cosmic radiation passing t h e crystals in transverse or longitudinal direction, respec tivel y.

d-

\

0)

C 3 0 0 -w

32 28 24 20 16 12 8 4

n

- 0 100 200 300 400

m,, MeV 3

Fig 10 Invariant mass spectrum for pairs of p r o m p t neutral hits in two coincident detector blocks. The dashed curve is a spline ^fitto the background.

5 Summary

The comissioning of the apparatus and the first performed experiments have shown t h a t the envisaged performance has been fully achieved. The strict specifications, the de- tailed performance tests, the careful assembly of the individual modules and the development of several instru- mental components optimized t o the needs of the TAPS- applications make i t possible t o install and operate such a highly segmented detector system a t several medium energy facilities. The detector performance based on t h e test measurements performed with prototypes makes it obvious t h a t the application of the spectrometer is not restricted t o a pure photon detection. The excellent response of BaFz t o charged particles (expressed bv energv resolution as well as particle identification clue t o pulse shape analysis) provides in combination with the CPV-system the basis for a high resolution AE-E detector system with the capability t o detect neutrons with high efficiency in the same device.

6 Acknowledgements

I would like to thank W. Doring for his support in test- ing and carefully assembling the detector modules. This work was supported in part by Deutsches Bundesminis- terium fur Forschung und Technologie and by Gesellschaft f i r Schwerionenforschung.

(7)

3200

28W

24M)

2rjm

1600

12w

ew

4 0 0

0 '

I ~ ~ ~ I ~ I I , . / I I I , , , , , , , , . . , , / , , /

0 400 600 1200 IS00 :WO

t o t a l light o u t p u t [ a.u. ]

Fig. 11

GeVlu 40Ar i natCa.

Pulse shape analysis performed in the reaction I

7 References

[l] Technical Proposal For A Two Arm Photon Spec- tometer (TAPS), GSI-Report 19, Nov. 1987

[2] N.Lava1 et al., Barium fluoride - inorganic scintillator for subnanosecond timing, Nucl. Instr. and Meth. 206 (1 983)169

[3] S.Majewski et al., G a m m a radiation induced damage effects in the transmission of barium fluoride and ce- sium fluoride fast crystal scintillators, Nucl. Instl,. and Meth. A&60(1987)373

[4] R.Novotny et al., Detection of hard photons with BaF2-scintillators, Nucl. InstT. and Meth.

A&62(198 7)340

[5] A.L.Boonstra et al.,Optical gain monitoring and cal- ibration system for TAPS, TAPS-Report l,?, April 1990, to be publashed

[GI N.Brurnmund et al., T h e charged particle veto detec- tor ( C P V ) for TAPS, TAPS-Report 14, May 1990, to be publashed

[7] R.Novotny, Performance of BaFz-crystals, TAPS- Report 10, Feb.1990 a n d TAPS-RepoTt 1 0 / l y Oct.

1990, znternal report

[8] 8-fold CAMAC-controlled modules: FFC'8' (Con- stant Fraction Discriminator) - RDV' (multiple gate generator, provides two independent gate outputs per input which are necessary for pulse shape analysis

(short and wide integration gate)) - QDC' (high resolution charge sensitive analog-to-digital-converter, optimized for pulse shape analysis by allowing two individual gates per analog input channel) - TDC' (high resolutioii time-to-digital converter)

16-fold ________ NIM module: Active analog fan-out ( 4 outputs of equal amplitude per analog input ) ['available from GANELEC]

[9] R.Novotny, Specificati0ns:Barium fluoride scintillator crystals for the TAPS photon spectrometer, T A P S - Report 7, Oct. 1988, anternal report

[IO] P.Schotanus et al.,The effect of Pb2+ contamination on BaFzscintillation characteristics, Nucl. Instr. and Meth. A272(1988)917

[I13 K.Wisshak et al., The Karlsruhe 47r barium fluoride detector Nucl. Instr. and Meth. A292(1990)595 [12] 0.Schwalb et al.,Test of a TAPS sub-array with elec-

trons, Nucl. Instr. and Meth. A&95(1990)191 [13] W.Arnold et al.,Production of a positron beam in

the energy range 1-10 MeV. Nucl. Instr. and Meth.

A256(1987)189

141 R.Brun et al., GEANT3, CERN/DD/ee/84-lJl 986 151 M.Appenheimer, The response function of a TAPS

sub-array t o high energy photons, Diploma thesas, Gtej3en 1990, to be publashed

161 A.L.Boonstra et al., Results test experiments Gronin- gen and Gieflen 1989, TAPS-Report 11, Aprd 1990, t o be publashed

[17] T.Matulewicz et al., Response of BaF2,CsI(Tl) and Pb-glass detectors t o neutrons below 22 MeV, Nucl.

Instr. and Meth. A274(1989)501

[18] T.Schick,Detection of neutrons with BaFz-detectors in the energy range froin 200 MeV t o 850 MeV, Daploma thesas, Gaepen, 1989, to be publzshed [19] F.D.Berg et al., Neutral Pion Production a t

350 MeV/u in the system "Ne

+

27Al, Short Note i n Z.Phys.A 9 1 7 ( 1 9 9 0 ) 1 5 1