Experimental Techniques in Nuclear Physics 9783110809824, 9783110144673

Citation preview

Experimental Techniques in Nuclear Physics

Experimental Techniques in Nuclear Physics Edited by Dorin N. Poenaru and Walter Greiner

W G DE

Walter de Gruyter Berlin · New York

1997

Editors Professor Dr. D o r i n N . Poenaru Institute o f A t o m i c Physics P.O. B o x M G - 6 R O - 7 6 9 0 0 Bucharest-Mägurele Romania email: [email protected]

Professor Dr. Dr. h. c. mult. Walter Greiner Institute o f Theoretical Physics J. W. G o e t h e University Postfach 111932 D - 6 0 0 5 4 Frankfurt a m Main Germany

This book contains 279 figures.

Library of Congress Cataloging-in-Publication

Data

Experimental techniques in nuclear physics / edited by Dorin N. Poenaru and Walter Greiner. p. cm. Includes bibliographical references and index. ISBN 3-11-014467-0 (alk. paper) 1. Nuclear counters. 2. Nuclear physics-Experiments— Technique. I. Poenaru, D. Ν. II. Greiner, Walter, 1935— QC787.C6E97 1997 539.7'7-dc21 97-19524 CIP

Die Deutsche Bibliothek — Cataloging-in-Publication

Data

Experimental techniques in nuclear physics / ed. by Dorin N. Poenaru and Walter Greiner. — Berlin ; New York : de Gruyter, 1997 ISBN 3-11-014467-0

© Copyright 1997 by Walter de Gruyter & Co., D-10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Printed in Germany. Typesetting: Thomson Press (India) Ltd., New Delhi. Printing: Ratzlow-Druck, Berlin. Binding: Lüderitz & Bauer, Berlin. Cover design: Hansbernd Lindemann, Berlin. Cover illustration by kind permission of GSI Darmstadt.

Contents

Introduction

1

1

Gas Filled Detectors Claude Stephan

11

2

Scintillation Detectors Klaus D. Hildebrand

59

3

Semiconductor Detectors J. P. Ponpon

87

4

Track Detectors and Soleno applied for Cluster R a d i o a c t i v i t i e s . . . . E. Hourany, I. H. Plonski and D. N. Poenaru

117

5

New Generation of Gamma-Detector Arrays Rainer Μ. Lieder

137

6

Modern Electron and Positron Spectrometers — Selected Aspects of a Broad Field H. Bokemeyer, J. van Klinken and P. Salabura

189

7

Neutron Detectors Gudmar Grosshög

235

8

Neutrino Detectors J. L. Vuilleumier, G. Zacek

271

9

Fragment Multi-Detector for the Study of Hot and Dense Nuclear Matter Produced in Relativistic Heavy-Ion Reactions J. P. Coffin

309

10 Production and Use of Radioactive Beams Isao Tanihata

343

11 In-Flight Separation of Heavy Ion Beams Gottfried Münzenberg

375

VI

Contents

12 The Measurement of Nuclear Lifetimes Klaus-Peter Lieb

425

13 Particle Identification Using Detector Telescopes Frangoise Pougheon

489

14 Particle Decay Study of Oriented Nuclei J. Wouters

497

15 Fission Fragment Mass, Charge and Energy Distributions Hans Otto Denschlag

535

16 Statistical Fluctuations in Nuclear Processes Mircea Penfia

583

Authors' and Editors' Addresses

637

Index

641

Introduction

The sixteen chapters of the book are briefly presented in the following few pages. Gas-filled detectors are still competing advantageously with other kinds of detectors. The first chapter starts with a short review of basic processes involved: charged particle energy loss; photon interactions; charge collection, and different operating regimes including the self-quenching streamer (SQS) mode. The large variety of detectors are grouped according to the operating regime. Various designs of ionization chambers are presented: gridded; transmission; position sensitive; Brag, and direct current chambers. The next section deals with the position sensitive, multiwire proportional counters, SQS mode detectors, drift chambers and pictorial drift chambers. Among the low pressure gas detectors one has to mention the parallel plate avalanche chambers as well as the low pressure multiwire proportional counters and the multi-step detectors. Finally the recently invented microstrip gas detectors, made by using microelectronic technologies, are described. They will certainly constitute an important development in the near future. Scintillation detectors are also very widespread. Stimulated by some medical applications as X-ray and positron emission tomography, new materials have been introduced. After a short presentation of the physical mechanisms by which the photons are emitted as a result of interactions of nuclear radiation with matter, many examples of commonly employed inorganic compounds are described, mentioning their advantages, disadvantages and area of utilization. One has for instance Nal(Tl), CsI(Tl), BaF 2 , B i 4 G e 3 0 1 2 (known as BGO), CaF 2 (Eu), and ZnS(Ag). In organic materials the scintillation is based on the excitation of individual molecules; it is not an effect of the crystal lattice. The two components of the obtained signal may be used to discriminate between γ-rays and neutrons. Three classes of materials are reviewed: pure crystals (anthracene and stilben); plastic scintillators, made from a primary agent dissolved together with a wave-length shifter in a polymerizing solvent (e.g. vinyltoluene, styrene, acrylic glass), and loaded scintillators with Pb, Sn or a material (like Β or Gd) with large cross section for neutron capture and subsequent charged-particle or γ-emission. Other kinds of scintillators are made from glasses doped with certain activators and from liquified noble gases (large volume calorimeters in high-energy physics). A special section is devoted to the light detection (photomultiplier tubes (PMT) and semiconductor photodiodes, optical coupling and sensitivity matching), energy and time measurements, and particle identification methods. High performance semiconductor detectors are available as a result of a continuous search for new materials and fabrication of devices adapted to a wide range of experimental conditions. For instance the advanced projects in the field of γ-ray spectroscopy are using large volumes of high purity germanium (HP-Ge) detectors. Starting from the two-dimensional microstrip detectors used for particle tracking in high energy physics, there is a trend of extending the microelectronic technology towards integration of the detector and the first electronic amplification stage on the

2

Introduction

same chip (the smart pixel detector system). The sensitive region and the charge carrier collection process of reverse biased planar or coaxial junctions, together with a detailed analysis of the need for various materials, are presented in the section on basic considerations. The next section deals with germanium detectors: lithium-drifted; γ-compensated; high purity; their timing properties; radiation damage and temperature behaviour. Then the silicon detectors are described: surface barriers and p-n junctions; lithium-drifted; position sensitive (including microstrip and conventional charged-coupled) devices; drift chambers; monolithic smart pixel detectors; amorphous material; radiation damage and annealing, etc. Some exotic materials are also mentioned: cadmium telluride; mercuric iodide; diamond, and SiC. Several solutions, e.g. the microgravity crystal growth in space have been envisaged to prepare high quality samples. In 1981, the editors together with A. Sandulescu predicted cluster radioactivities. Since 1984 many successful experiments have been performed. The collected data on the half-lives of trans-francium nuclei against 14 C, 20 O, 23 F, 24 ~ 26 Ne, 28 ' 30 Mg, and 32,34 Si decay modes are in good agreement with our predictions based on the analytical superasymmetric fission model. A fine structure in the 14 C decay of 2 2 3 Ra, discussed by Martin Greiner and Werner Scheid in 1986, was experimentally discovered in 1989 by E. Hourany, M. Hussonnois and their colleagues at IPN Orsay, by using the superconducting magnetic spectrometer SOLENO, an instrument with extremely high performances (large solid angle, high energy and mass resolution, and good rejection of the extremely high background of α-particles). SOLENO had already been used in 1984 to confirm the first experiment carried out by Rose and Jones with a standard AE χ Ε telescope in direct view of the source. They had to run half a year in order to get 11 events, which have been obtained in 5 days with SOLENO. The construction, the working principles and the main characteristics of the spectrometer as well as preparation by implantation of strong sources of high uniformity, the detection setup, and some typical experiments on 14 C decay and its fine structure are presented. Another technique, very frequently used in Berkeley (Price et al), Dubna (Tretyakova et al), and Milano (Bonetti et al), allowing branching ratios as low as 10" 17 to be measured, is based on the solid state nuclear track detectors (SSNTD), which are not sensitive to alphas and other low-Z particles. In a series of materials with good insulating properties, as for instance some plastic foils (polycarbonates, polyethylene terephthalates), mica, quartz, phosphate glasses, etc, the ionizing particles produce a latent track. This is a narrow region around the trajectory in which the material is damaged. After a subsequent etching with a chemical reagent the track is enhanced and can be seen with an optical microscope. The related topics are: track formation, detector characteristics and materials, the background and loss of sensitivity due to a long-term exposure to alpha particles, particle identification and determination of its kinetic energy from the geometrical parameters of the track (measured after a manual or automated scanning of the detector surface), calibration with a beam of accelerated heavy ions, etc. Finally some examples of experiments on cluster radioactivities, as well as other applications in experimental nuclear physics (fission, identification of heavy elements, search for superheavies, reaction mechanisms, measurements of very short half-lives by "shadow effect", etc.) or in technology of special micropore filters (for pharamaceutical, milk and wine industries or ultraclean workshops making micro-

Introduction

3

electronics devices) are described or mentioned. The illustrating examples are mainly from cluster decay modes; extension to other applications can be easily made. The energy resolution of gamma-ray detectors was improved in the late 1960s by introducing germanium detectors. Their peak-to-total ratio was increased in 1980 by using escape-suppression scintillation (first Nal(Tl) and recently BGO) detectors which allow the Compton background to be reduced. The most recent progress, of multiplying the total photopeak efficiency by a factor larger than 50, has been made by increasing the number of escape-suppressed Ge detectors (up to 54) in an array surrounding the target. The discovery in 1986 of the superdeformed bands at high spins by Twin et al. has been possible on the basis of this technological progress. Even larger arrays (EUROBALL III in Europe and G A M M A S P H E R E in USA) have been constructed. High spin states in nuclei are efficiently populated by fusion reactions. The decay of compound nucleus proceeds by neutron evaporation. Each neutron takes away about 8-10 MeV from the excitation energy and ~ 1 h of angular momentum. Gamma-decay starts when the excitation energy above the Yrast line is less than the neutron separation. Electric dipole transitions are again removing much excitation energy and small angular momentum. The stretched electric quadrupolar transitions (E2) are important at lower excitations above Yrast line. A discrete γ-ray spectrum appears on top of a background formed by the statistical decay and the quasicontinuum. It is explained how one can establish the level scheme of a nucleus (energy, spin, parity, mean life-time, static and dynamic electromagnetic moments, of each level) in experiments of in-beam γ-spectroscopy. The parameters characterizing the performances of a γ-detector array (peak-to-background ratio and the importance of high-fold coincidences, energy resolution and Doppler broadening, total photopeak efficiency and peak-to-total ratio and other features (e.g. granularity)) are defined and discussed. The escape-suppressed Ge, the composite (CLOVER and CLUSTER which can be used as Compton Polarimeters) and the segmented detectors, are presented in the section on high resolution spectrometers. Three large instruments available for experiments at present (GASP, EUROGAM II, and G A M M A S P H E R E I) and the next generation (EUROBALL III and the full GAMMASPHERE) illustrate the high performances one can achieve with these complex detector arrays. The new concept of digital pulse processing will be used in the associated electronics data acquisition system. Electrons and positrons are present in a large diversity of atomic and nuclear phenomena, with an energy spectrum covering an extremely wide range (from meV to GeV). Spontaneous positron emission from quasi-atoms with very large Ζ (over 172), produced in heavy-ion collisions, could be the signature of the decay of neutral vacuum into a locally charged one, predicted by W. Greiner and his coworkers in their study of electrodynamics of strong fields. The devices described in the chapter on modern electron and positron spectroscopy are usually designed for low intensities and cover an energy range from about 100 keV to a few GeV. One has to take into account the physical processes (collisions with atomic electrons, bremsstrahlung and electromagnetic showering) of slowing down through matter, as well as the Doppler shifts owing to fast moving emitters. Particular attention is paid to different sources of background (external and internal pair conversion of γ-rays, knock-on electrons, (5-electron emission, and sources produced in relativistic heavy-ion collisions: π + -decay, dilepton

4

Introduction

decay of ρ and ω vector mesons, etc.), which can sometimes be some orders of magnitude larger than the studied process. One section is devoted to the motion of electrons and positrons in solenoidal and toroidal magnetic fields used in spectrometers. Three families of instruments are presented in the next section: transportfield spectrometers (for conversion electrons, for positrons, and for electron positron pairs) using Si(Li) or Si PIN diodes, dispersive spectrometer devices (orange and mini-orange types), and trajectory spectrometers for very high energies. The transport systems (solenoidal and toroidal fields as well as dispersive orange geometries) allow at least an important part of the background to be suppressed. The recoil shadow technique eliminates the promptly emitted ^-electrons. The positrons are more difficult to measure owing to their lower (by about four orders of magnitude) yield. They can be selected by means of helical baffles and identified by detecting the annihilation radiation in a Nal(Tl) scintillator surrounding the Si(Li) detector. Alternatively, a S-shaped solenoid may be used to separate spatially the electrons and positrons by the gradient shift method. With a 4π pair spectrometer (e.g. EPOS II and APEX) one can simultaneously measure coincident pairs, after very heavy ion collisions, at all lepton emission angles. A very good collection efficiency and a Doppler shift correction can be achieved with orange and mini-orange devices using toroidal magnetic fields. The most advanced instrument of this kind is the double-orange spectrometer, used for in-beam positron and electron spectroscopy, which provided outstanding background suppression possibilities. The Bhabha scattering spectrometer RESBHA and the 4π dilepton spectrometer PEPSI are examples of complex mini-orange pair instruments. At very high lepton energies, beyond 100 MeV, the tracking techniques with multi-wire gas detectors, Si strip detectors, or drift detectors, are suitable. A dilepton spectrometer in this energy regime combines lepton identification by Cerenkov detectors with tracking in drift chambers. The new systems HADES now in construction, will be used in heavy ion collisions at incident energies of the order of 1 GeV/u. Proton scattering or nuclear reactions on H, 10B, 6 Li, 3,4 He, or fissionable materials ( U and 2 3 9 Pu for thermal neutrons, and 238 U, 2 3 7 Np and 2 3 2 Th for fast neutrons) are produced in the first stage of a neutron detection in order to obtain ionizing particles. Each reaction is separately analysed. The simplest neutron detectors are the neutron counters, allowing the number of neutrons to be ascertained. The active material is included in a ionization chamber, either as a part of the counting gas (e.g. BF3), or (for instance in a fission chamber) as a coating layer on the wall. Many neutron scintillation detectors are available. One may use proton recoils in organic materials, lithium or cesium loaded glasses, boron loaded liquid scintillators, combination of boron plates with Nal(Tl) scintillators, optical fibres composed of plastic scintillating materials, barium fluoride, cesium iodide, lead glasses, etc. A combination of the commercial NE213 and a CaF 2 (Eu) connected to the same PMT can be used to discriminate n, p, d, t, α in the energy range 10-100 MeV. Neutron moderators (e.g. paraffin wax or polythene) enable the sensitivity of the fast neutron counters to be improved. Moderators and absorbers can be arranged to obtain a response function simulating that of a human body (REM counter). Neutron spectrometers are generally more sophisticated devices, based on elastic proton scattering or 3 He reaction. A resolution of 20keV for 0.5-5 MeV neutrons can be achieved with a gridded ionization chamber. 3 He is also employed in a 1 mm space between two semiconductor detectors 235

Introduction

5

(the sandwich detector), or as a mixture with Xe in a gas scintillator. The best energy resolution is obtained with the time-of-flight (TOF) spectrometer. Other types of neutron spectrometers are: 4π proton recoil spectrometer, the proton recoil telescope and the double scattering spectrometers (at the Joint European Torus fusion facility). The main sources of low energy neutrinos are: β ~ activity of fission fragments in high power nuclear reactors; solar thermonuclear fusion processes (p-p chain and C N O cycle); stellar collapse leading to a neutron star in supernova explosions, and the capture of atomic electrons in strong (1 MCi) sources o f 5 1 Cr or other isotopes a s 3 7 Ar, 62 Zn and 1 5 2 Eu. The existence of a non-vanishing neutrino mass, of its magnetic moment and of neutrino oscillations (modulation of a monoenergetic beam intensity along its direction of propagation due to a possible change of neutrino flavour ve, νμ, ντ (non-conservation of the lepton quantum number)) are some of the important problems of neutrino physics to be solved in the near future. These weak interaction particles are detected on the basis of several reactions like: neutrino capture on heavy nuclei; inverse ß-decay of the neutron; neutral and charged current scattering on deuteron (in a large tank of heavy-water); neutrino electron scattering, etc. The resulting rare events are recorded by huge volume detectors (e.g. Li —, Cd — or Gd loaded liquid scintillators, 3 He multiwire proportional counters, etc.). Various experiments performed to date at some high power reactors (Hanford, Savannah River, Gösgen, Rovna and Bugey) are described. Three of the four successful experiments performed so far ( 37 C1 and 7 1 G a (GALLEX and SAGE)) on solar neutrino detection are based on radiochemical methods (looking for some nuclei activated by neutrino interactions after a long-time exposure) and one (Kamiokande) by direct measurement technique with a water Cerenkov detector. The future projects of solar and reactor neutrino experiments will use Cerenkov detectors (Superkamiokande, Sudbury), boron loaded scintillation detector (BOREXINO) or a liquid argon ionization detector (ICARUS). The production of hot and dense nuclear matter in reactions of relativistic heavy ions (100 to 600 A MeV) has been proposed by Greiner and coworkers since 1969. A strongly compressed state of the nuclear system is created by the propagation of a shock wave from the impact zone of colliding nuclei. The decay of this state is a complex phenomenon called multifragmentation and finally leads to generation of light particles (n, p, d, t, 3 ' 4 He) and fragments with low mass numbers (A = 4-30), which have to be recorded event by event over the full solid angle. One has to identify (A and Z) each particle and to measure its velocity or linear momentum (including the polar and azimuthal angle of emission). From the primary variables, associated quantities are deduced: multiplicity; rapidity; kinetic energy, etc. The F O P I (4π) facility installed at G. S. I. Darmstadt is described in detail as a comprehensive and illustrative example. The basic requirements, concepts and solutions of the detector design are analysed in detail. In this way the final solution with its two phases (one operational since mid 1990 and the other since the end of 1993) emerges naturally. The first phase is devoted exclusively to multifragmentation detection in a coverage of polar angles in the laboratory system 0° θ ^ 30°. In the second phase the solid angle is extended and low Ζ (1 and 2) particles are also detected. In fact the main emphasis is put on π ± and Κ meson detection. The overall size is about 4 m diameter and 3 m depth in the first phase. The target and the detectors are placed in air. The plastic scintillation detectors forming the inner (252 units) and outer wall (512 strips) measure the energy loss AE

6

Introduction

and the TOF, hence the velocity, to allow Ζ identification (from Ζ = 1 to 9). The cluster detector system comprises the "Rosace" (60 thin plastic scintillators) and the "Parabola" (16 gas filled ionization chambers) allowing the particle identification to be extended up to Ζ = 18 and lower rapidity events to be measured. One also has to mention the start detector, halo-supressors and the reaction counter system. All scintillation modules are connected by quartz fibers to a single N 2 laser calibration system. The associated electronics located in the counting room are connected by a 45 m cable to detectors. The accumulated statistics during a run are monitored and recorded onto 8 mm video cassettes for off-line analysis. In the second phase a detection ensemble placed in the magnetic field of a superconducting solenoid (3.8 m long with a diameter of 2.4 m) is added. The central drift chamber with 16 radial segments identifies particle mass by means of dE/dx versus momentum (from Bp) analysis. A plastic BARREL of 180 scintillator strips yields T O F measurements hence an additional mass determination. An equal number of water Cerenkov detectors allow a good Κ/π discrimination. There is also the H E L I T R O N system of 24 radial drift chambers. At present this instrument gives the best performance in its class. The recent development of radioactive beam facilities provides new opportunities to extend the production and study of nuclear species far from beta stability. Close to the neutron-drip line, nuclei (like 11 Li) might exhibit a halo structure or neutron-skin. In these nuclei mass and charge radii may differ by large amounts, and the molecular (or cluster) structure plays an important role. The density distributions show an extended tail at low density. On the neutron-rich side, it was observed that pairing is weakened for large values of N-Z. Nuclear physics is also closely linked to astrophysics, cosmology and elementary particle physics, helping to understand the unity of our knowledge about the universe. The main site of nucleosynthesis for most of the elements is considered to be the stellar environment; a second is the Big Bang (primordial nucleosynthesis). In both cases, reactions of ß-unstable nuclei play an important role. Nuclear reactions involving short-lived nuclei in explosive processes in the universe could be simulated in the laboratory. Two main production methods are discussed: the secondary beam (projectile fragmentation in a high energy heavy ion reaction, followed by a high efficiency recoil mass separation based on the magnetic rigidity (A/Z) and on the energy loss into a degrader (Z 2 )) and the reacceleration. In the latter case the ISOL facilities use spallation or fission reactions induced by high energy protons. Difficulties appear in the extraction stage which is inherently long, hence short-lived isotopes have no chance to survive. Specific examples of both kinds are presented in operation at LISE (GANIL), RIPS (RIKEN), FRS (GSI), Louvain-laNeuve, Notre-Dame. Also some new projects (ISOLAB in USA and J H P in Japan) are being developed. The above mentioned studies of nuclear structure and nuclear astrophysics are presented in detail. Also possible applications in cancer therapy and for other purposes are mentioned. New refinements have been recently added to various kinds of recoil separators. In this way extremely high performances have been reached. The lowest rate in a reliable identification of an isotope was extended down to one atom per week. Energetic secondary beams are produced by neutron induced fission (neutron-rich nuclei), proton induced spallation of heavy nuclei, complete fusion of heavy ions (proton-rich nuclei), nuclear fragmentation of relativistic heavy ions, etc. Reaction kinematics and

Introduction

7

beam luminosity are discussed. An important section is devoted to the ion optics (allowing raytrace calculations to be made and the optimum positions of the second order correcting elements providing high resolving power to be found) and design principles of the in-flight electromagnetic separators. They are able to select a given mass to charge ratio. In order to achieve isotopic resolution one has to include the energy loss analysis in the beam degraders. The main difficulty one has to surmount is the adaptation to the reaction kinematics. The recoil products of a fusion reaction possess a whole distribution of charge states unlike the relativistic ions which in general are totally stripped. Several examples of in-flight separators are given: recoil mass separators (Osaka, Rochester, Legnaro, Oak Ridge, Argonne); velocity filters (SHIP at GSI used to detect the first ground state proton emitter and to produce the heaviest elements 107-112, the Daresbury separator with crossed-field velocity filters); radiofrequency (RF) separators (TOFI at LAMPF, Los Alamos, combining T O F with magnetic momentum analysis and the velocity filter of the München separator combined with T O F analysis using an electric RF deflection field); parabola spectrographs (LOHENGRIN at ILL Grenoble) and gas filled separators able to increase the transmission even for low phasespace densities (SASSY in Berkeley, GARIS in Tokyo, NASE in Darmstadt, RITU in Jyvaskyla). Finally the separation of relativistic heavy ions are illustrated by describing the facilities installed at the following research centres: GSI Darmstadt; BEVALAC Berkeley; LISE at GANIL Caen; RIPS at RIKEN Tokyo and COMBAS at JINR Dubna. A new generation of experiments will benefit from the availability of cooled secondary beams in storage rings allowing a wide range of energies with ultraclean separation to be spanned and at the same time high momentum resolution and brilliance. The experimentally accessible range of nuclear lifetimes covers more than 45 orders of magnitude. Specific methods have been designed for different subdomains as for instance very long lifetimes in α-decay or very small activities of exotic nuclei produced in heavy ion fusion reactions, thermal neutron induced fission, or relativistic heavy ion fragmentations. The neutron lifetime has been measured with high precision by employing in-beam and ultracold neutron storage (magnetic bottle or material wells) methods. A large variety of direct timing methods are available in the range of lifetimes (10~ 15 , 10~ 6 )s. The traditional fast-slow coincidence circuit is up-dated by new scintillation detectors (e.g. BaF 2 ) and faster electronics; it can reach a lower limit of lifetimes as short as 10 ps. The fast and compact detector for positron tomography uses Ge-arrays with multi-element Compton suppression (CsF, BaF 2 and BGO). Shortlived fissioning shape-isomers have been identified and the lifetimes of rotational states in the second minimum of the fission barrier have been measured by the recoil shadow timing method with a special charge-plunger device. Also a very sophisticated plunger is used in the recoil distance plunger method applied over the whole periodic table for the excited y-decaying nuclear states. Corrections due to reaction kinematics, delayed feedings, continuum or side-feeding radiation and hyperfine deorientation are made in the analysis of this kind of experiment. The Doppler shift attenuation method (in the range 5· 10" 1 5 to 3· 10" 1 3 s) - a standard technique in nuclear spectroscopy - exploits the stopping time of energetic ions in materials as a clock for γ-emitting nuclear states. The lifetime is deduced from the broadened line profile. With the gamma-ray induced Doppler broadening (GRID) method one measures short lifetimes of states populated

8

Introduction

in thermal neutron capture reactions. There are no generally applicable direct timing techniques in the range 10" 1 7 to 10~ 1 5 s. For ß-delayed proton emission the X-ray timing method has been developed, based on the finite halflife of K-shell vacancies. Crystal blocking is used to measure compound nuclei lifetimes. The recent interest in the scissor-mode vibrations has revived the field of methods based on nuclear resonance fluorescence (Mössbauer effect). The problem here is to find a suitable photon source. One may use either a high flux fission reactor with an internal target or the bremsstrahlung radiation produced by an intense electron beam hitting a heavy nucleus target. For the light-element production through low-energy resonances, of importance for astrophysics, the nuclear reaction analysis allows the narrow resonance widths to be determined. The particle identification using the AE-E detector telescope is based on a simplified version of the Bethe-Bloch relationship for the energy loss in matter, according to which the product of signals from AE and Ε detectors is proportional to the square of the charge number Z 2 . The mass number can be also determined by measuring the T O F along a well known flying path. The doubly achromatic LISE3 sepctrometer at GANIL combines a magnetic spectrometer with a Wien filter. It is also combined with an achromatic degrader in the intermediate focal plane. It was recently improved by making the entrance angle of the primary beam variable with respect to the axis (allowing the incompletely stripped ions to be suppressed) and by installing a selection in velocity based on an electrostatic field crossed with a magnetic one. The flight path between the target and the final focus is now increased to 43 m. Many new proton-rich and neutron-rich nuclei have been produced and identified by using this facility. In order to observe the anisotropy of emission of quanta with nonzero angular momentum, the parent nuclear spins should possess a nuclear orientation (NO) in space. The static low temperature method of N O is based on the hyperfine interactions of nucleus with its surroundings. Recently N O in combination with a positron Polarimeter has been used to provide constraints on extensions of the Standard Model; new limits on right-handed charged weak currents have been established. In a combination of particle (α, β, fission fragments) detection with on-line N O at ISOL systems one can measure radiation rates down to about 100 particles/s (few counts per s in detector). A larger detection solid angle would average out the anisotropy. In order to achieve the NO, the radioactive species should be brought into selected host materials like a ferromagnetic foil with internal fields as high as 10-100 T. A dilution 3 H e - 4 H e refrigerator allows the required low temperature (5 mK) to be obtained. The host foil is magnetically saturated with an external field of about 1 Τ provided by a superconducting solenoid. In a on-line setup, like that installed in 1981 at Lovain-la-Neuve, the wanted isotope produced by nuclear reactions is separated by an ISOL facility, transported and implanted in the ferromagnetic foil where it achieves the NO. Similar facilities are also available at DOLIS in Daresbury, UNISOR in Oak Ridge, and ISOLDE-CERN in Geneva. The detection system should be mounted inside the cryogenic environment. Good resolution and detection efficiency has been obtained with HP-Ge and η-type based Si junctions. Typical resolutions of 4 keV for 1 MeV conversion electrons and 16 keV for 5 MeV α-particles are not difficult to obtain. The measurements performed at Columbia University on parity non-conservation in weak interactions are described. The beta-emitter source (e.g. 6 0 Co) is prepared by ion

Introduction

9

implantation and oriented in cerium magnesium nitrate. Nuclear magnetic resonance (NMR) of oriented nuclei is now a standard method used to study solid state parameters related to hyperfine interactions and magnetic structure. The anisotropic alpha decay is presented in detail. As early as 1953 Hill and Wheeler predicted the effect of the non-spherical shape of an α-emitter on angular distribution of emitted particles, leading to a preferential emission from the poles (parallel with nuclear spin) of a prolate parent nucleus (where the potential barrier is lower and thinner). These qualitative conclusions are confirmed by the experiments. Close to the magic neutron number of neutrons Ν = 126 the α-particles are emitted preferentially parallel to the direction of nuclear spin, whereas they are emitted mostly perpendicular to this direction at some lower neutron numbers. A detailed theoretical approach is still missing. The neutroninduced fission of odd mass heavy nuclei (e.g. 2 3 3 · 2 3 5 υ and 2 3 7 Np) is accessible for N O studies. The anisotropic distribution of fission fragments strongly depends on the projection quantum number K. Besides the low temperature orientation method there are other techniques available as for instance the collinear laser spectroscopy, the tilted foil technique, reflection from very flat Si single crystal at grazing incidence, etc. Nuclear fission is certainly the most complex decay process, in which more than 500 different nuclides are produced. Fission fragment mass, charge and energy distributions not only provide basic information on this large amplitude collective motion, but they are also important for certain applications such as nuclear power plants, safeguards, and nuclear incineration. The various yields in use (independent, cumulative, chain, fragment, mass number yield, etc.) are defined in the introduction of the corresponding chapter. The methods of measurements are generally based on: radiochemistry and classical mass-spectrometry; direct γ-ray spectroscopy; fission fragment energy and T O F determination, etc. The corrections for the prompt neutron emission and the pulse height defect in Si detectors allow the distribution of primary fragments to be deduced prior to neutron emission. The main facilities are: HIAWATHA, COSI FAN TUTTE, and the mass separator of fission products L O H E N GRIN. Typical results on mass distributions show the dependence of mass asymmetry on fissility parameter and excitation energy. By selecting the total kinetic energy (TKE) of the fragments, one can derive the contribution of cold fission process in which the released energy is almost exhausted by TKE. An increased sensitivity in recently performed experiments allowed new neutron-rich nuclei to be identified as fragments of very high asymmetry. Some calculations of mass distribution are compared with data. By combining Ζ identificaion by specific energy loss with a mass separator, one can determine simultaneously the distributions of masses, charges and kinetic energies of fission fragments. The results on charge distribution are discussed in terms of model descriptions and various kinds of odd-even effects. The total kinetic energy of fission fragments increases with the fissility parameter from about 160 MeV for 2 3 2 Th to 270 MeV for the heaviest elements. If we ignore the measurements on cold fission processes, it is generally rather well described by a semiempirical relationship known as Viola systematics. Any experimentally determined quantity should be accompanied by an estimated error also expressing the confidence we can have in the final result. Except for the systematic errors which are a kind of mistake, the statistical nature of the physical processes and the finite precision (or the uncertainty) of our instruments are reflected in

10

Introduction

some inherently present and reproducible fluctuations of measurements around the "true" mean-value, which are generally called statistical fluctuations and random errors, respectively. The random errors could in principle be reduced by using more precise measuring instruments and the statistical ones by counting more events. The related statistical populations are characterized by a probability density of the random variable, which in turn is determined by a small number of parameters like the mean value and the variance. Some useful derived quantities are the deviation and the standard (or rms) deviation. The variance is a measure of the spread of data around the mean value and the standard deviation gives the uncertainty of the final result due to the fluctuations in observations. When our measured quantity is a complex one, depending on several independent variables individually determined, one can speak of the propagation of errors and of the covariance, which is the two-variable analogous of the one-variable variance. The probability density functions commonly encountered in physics applications are the following distributions: uniform; binomial; Poisson; Gaussian (or normal); Breit-Wigner, and χ2. For each one the probability density function, the mean-value, and the variance is given. For a Gaussian distribution there is a simple relationship between the full width at the half maximum and the standard deviation. The confidence level defined for a χ2 distribution is a measure of the exactness of a fit. In the data analysis one is usually concerned either with hypothesis testing (to check whether the data are consistent with a given assumption) or the parameter estimation (the obtained data are used to determine the parameters of a model). The method of moments and that of the maximum likelihood can be employed to estimate a parameter. The first moment about zero is the expectation or the mean value and the second moment about the mean is the variance. Very frequently the obtained data versus a given independent variable are fitted, by the method of least squares, to a given simple law like: a straight line, a second degree polynomial with constant or statistical uncertainty, a η-degree polynomial, or to an arbitrary nonlinear function. A subject index is provided at the end of the book to facilitate reference. Dorin N. Poenaru and Walter Greiner

1

Gas Filled Detectors

Claude

St0phan

Table of Contents I. Introduction 12 II. Basic processes involved 12 A. Energy loss in gas 13 1. Light particle energy loss 13 2. Heavy ion energy loss 13 3. Photon interactions 14 B. The relation between energy deposited and charge collection 1. Number of ion pairs created 16 2. Ion drift 18 3. Electron drift 18 4. Recombination 19 5. The multiplication phenomenon 19 6. Different regions of operation of gas counters 20 III. Ionization chambers 21 A. Principle of operation 21 1. The gridded ionization chamber 22 2. Wall and window effects 23 3. Observed signals 23 B. Various designs of ionization chambers 25 1. Transmission chambers 26 2. Position sensitive ionization chambers 27 3. Bragg chambers 28 4. Direct current ionization chambers 28 IV. Proportional counters 29 A. Principles of operation 29 1. Gas multiplication 29 2. Detected signal 31 B. Gas operation 31 1. Choice of gas filling 31 2. Gas purification 33 3. Gas pressure regulation 33 C. Various types of proportional counters 33 1. Position sensitive proportional counters 34 2. Multiwire proportional counters 36 3. Detectors operating in the self quenching streamer mode 4. Drift chambers 39 5. Pictorial drift chambers 41

12

Claude Stephan

V. Low pressure gas detectors 42 Α. Parallel plate avalanche chambers 42 Β. Low pressure multiwire proportional counters 45 C. Low pressure multi-step detectors 45 VI. Gas microstrip detectors 46 A. Description 46 B. Influence of the substrate and nature of electrodes 48 C. Performances 49 VII. Conclusion 51 References 52

I. Introduction Gas detectors are among the oldest and most widespread types of instruments used in nuclear physics. They enjoyed a return to popularity in the early 1970's with the acceleration of heavy ions which introduced new requirements: thin entrance windows, low pressure of operation, detectors set in vacuum, and chiefly the ability to work at the same time with particles with very different values of atomic number. They are radiation insensitive and they can compete advantageously with scintillators or solid state detectors which present pulse height defects increasing with Z, shortened life time due to large energy losses in the detector, and small active areas. Gas detectors are extremely versatile: they can handle high counting rates and be used for energy measurements, position determination or for timing with all types of charged particles, and even with γ and X rays. They can easily fit every experimental condition and be built at a relatively low price although in large dimensions, have the advantage of being able to be repaired on site during an experiment. Their disadvantages are essentially a relatively limited energy resolution and the necessity of a gas filling which might cause leak or window problems. The main types of gas detectors, ionization chambers, proportional counters, and avalanche counters, which have become routine tools, will be described. Complex designs and new devices are also advocated. For a better understanding of their way of operating, a resume on the processes involved in such detectors is given.

II. Basic processes involved The processes involved in gas detectors have been extensively studied and are now well understood. A charged particle traversing a gaseous medium essentially interacts with it by Coulomb interactions. The electromagnetic field of the particle will produce excitation and ionization when interacting with the outer electrons of the atoms of the gas. The energy loss results from these discrete interactions between the projectile and atoms from the medium. These primary collisions then create electron-positive ion pairs which again will be able to interact with the medium. The characteristics of the signals from the different types of gas counters are strongly related to the created

Gas Filled Detectors

13

number of primary pairs. In the absence of other effects, ions will rapidly lose their energy by multiple collisions with gas molecules while electrons will diffuse in the medium.

A. Energy loss in gas Many review articles have been written on the subject of charged particle energy loss in matter [1-2].

1. Light particle energy loss The energy loss of particles depends on the energy transferred to electrons of gas molecules during the interactions. The quantitative evaluation of the energy loss d Τ per length unit dX has been described a long time ago by Bethe and Bloch in the framework of relativistic quantum mechanics by the formula: dT_4jteV -nZ dX~ m0V:

2 m0V2 ln-ln(l I

—

β2) — β 1

Ζ

(1)

where Ζ is the atomic number of the medium, I is its effective ionization potential which can be taken with a good approximation as / = / 0 Z , where I 0 = 12 eV; η is the number of interactions per volume unit, m 0 and e are the electron mass and charge, Vand ζ are the velocity and charge state of projectile, β = V/c. C K represents inner shell corrections and can usually be neglected. In this equation (1) the projectile energy loss is a function of its velocity and not of its energy. The stopping power is then the product of two factors, one is a monotonically decreasing function of the velocity of the projectile, the second (between braces) increases logarithmically with it. This second term is responsible for the relativistic rise of the stopping power. If the energy is expressed in MeV and the thickness of the medium in gem ~ 2 one finds Λπ/72„7 7 2 =0.307j6~ 2 z 2 p —MeV/gcm~ 2 m0V A

(2)

A, p, are respectively the mass and density of the medium. For molecular gases or mixtures, one takes average values for A, Z, and I in the Bethe formula.

2. Heavy ion energy loss The Bethe formula (1) predicts with a good precision proton and alpha particle energy losses. Presently gas detectors are mostly used in nuclear physics for heavy ions. In order to apply the formula to these particles it is necessary to know the charge state ζ of

14

Claude Stephan

the projectile. Indeed it depends on the energy of the ion and on the traversed medium. It means that ζ will even vary during its path in the gas when decelerating. This renders in most cases the formula (1) almost unusable for heavy ions. As stopping powers and ranges play an important role in many aspects of heavy-ion physics, semi empirical tables of energy loss of heavy ions have then been established by means of a scaling law based on a compilation of hundreds of experimental data [3-5]. The most recent tables by Hubert et al. [5] have used the following parameterization equations where alpha particles are chosen as reference projectiles.

^ion

—

^He

(yuiZJi

2 ion

(3)

with yu2 = y J ( z „ z 2 ) y^l-yifZJexp

(4) -0.88F 0 65

(5)

Z j is the atomic number of the projectile, Z 2 of the m e d i u m . / ( Z ^ Z 2 ) is a function derived from a fit to all existing experimental data. Fis the projectile velocity, F0 is a constant, F0 = 2.18810 8 cm/s. These expressions give reliable results for all elements between 2.5 MeV/u and 500 MeV/u in a large sampling of solids. For gases, the authors find their tabulation valuable when the ions are fully stripped, but the energy loss of gases is decreased by 20% for ions which have a broad charge state distribution. This effect was first observed by Geissei et al. [6] in 1983. Fig. 1 shows such differences obtained by Herault et al. [7] for the energy loss of Xe ions in argon. The authors interpret the difference by a density effect in solids.

3. Photon interactions Gas detectors can also be used to detect γ and X rays. In the photon energy range considered in nuclear physics, i.e. from about 50 keV to about 50 MeV, three major types of interaction mechanisms lead to partial or complete transfer of the photon energy to ejected atomic electrons: photoelectric absorption, Compton scattering and pair production. In the photoelectric effect, the ejected atomic electron has a well defined energy line spectrum T ; given by the expression (6) Ti = hv — Wi

(6)

where Wt is the binding energy of the atomic level from which the electron was ejected. This well defined energy is commonly used in γ-ray spectroscopy. Photo-electrons can be ejected from any of the K, L, M , . . . shells of an atom resulting in discontinuities in γ-ray absorption which correspond to K, L, M , . . . energy levels. The efficiency of the

Gas Filled Detectors

15

Fig. 1. Variation of the stopping power of Argon for O, Ar, Kr and Xe ions versus incident particle energy. The experimental data (full dots) are compared with calculated curves from Ziegler (dotted lines) [4], from the "fully-stripped approximation" (solid lines) and from an effective charge parameterization valid for solid media (broken line) [5]. From Herault et al. [7]

photoelectric process, that is the cross section σφ, is dominant at low energy and depends largely on the charge Ζ of the absorbing material. Compton effect is due to an inelastic incoherent scattering of photons by the atomic electrons of the medium. This process becomes dominant for photon intermediate energies, large compared to the atomic binding energies. The kinetic energy of the recoil electron considered as a free electron depends on the scattering angles involved in the reaction. Consequently Compton electrons have a wide energy distribution spectrum ending at the incident photon energy hv. The cross section for Compton scattering ac has been carried out by quantum-mechanical treatment of the problem by Klein and Nishina [8]. All experimental tests of Compton scattering have shown agreement with theory. Pair production effect, that is the production of a positive and negative electron in the Coulomb field of the nucleus, becomes possible at photon energies above 2m 0 c 2 = 1.022 MeV. The energy observed is very close to this threshold, the recoil of the nucleus ensures the conservation of momentum. When the recoil is absorbed by an electron, the

16

Claude Stephan

threshold becomes 4 m 0 C 2 and two electrons and one positron are emitted with appreciable momentum. The pair production cross section is of the order of r\Z2j 137 where r0 = 2.82 χ 10" 1 3 cm is the classical radius of the electron. In contrast to charged particles which gradually lose their energy, here these processes result in sudden energy deposit in the gas by ejected atomic electrons. In order to get a high detection efficiency the choice of the gas filling will then be determined essentially by its absorption capability I, g-(AiMK^ + »f + t,)x

(1)

Here Ν is Avogadro's number, A is the atomic weight, χ the absorber thickness.

B. The relation between energy deposited and charge collection 1. Number of ion pairs created Regardless of the detailed mechanisms involved, the practical quantity of interest is the total number of ion pairs created along the track of the radiation. The minimum energy I 0 which can be transferred for each interaction is between 10 and 20 eV, corresponding to the ejection of the least tightly bound electrons. However, other mechanisms like atom excitation which happen in distant collisions can lead to an energy loss of the incoming particle without removal of an electron from the gas molecule. Therefore, the average energy W which corresponds to an ion pair creation is substantially greater than the ionization energy. The parameter Wdoes not depend much on the particle, its energy or even the nature of the gas. This feature is illustrated in Table 1 where values of W f o r alpha particles, 340 MeV protons and β rays are given [1,9],

Table 1. Effective ionization potential /0(eV) and values of W^eV/ion pair) in various gases for α, β rays and 340 MeV protons. From FANO [1] and from SAULI [9] where references are given Gases He Ne Ar Kr Xe H2 N2 O2 co2 CH4 C 4 H 10

I0

24.6 21.6 15.8 14.0 12.1 15.4 15.5 12.2 13.7 13.1 10.8

W

α rays

Β rays

42.7 36.8 26.4 24.1 21.9 36.3 36.4 32.5 34.0 29.0

42.3 36.6 26.4 24.2 22.0 36.3 34.9 30.9 32.9 27.3

340 MeV protons

36.5 34.7 32.6 23

Gas Filled Detectors

17

F r o m T a b l e 1 one can deduce that a 10 M e V particle will create about 3· 105 ion pairs. There are fluctuations in the number of these ion pairs and these fluctuations can have important consequences for the energy resolution of the detector. It is necessary to distinguish t w o cases: when the particle deposits a large amount of energy in the detector or when it leaves only a small part of its total energy. In the first case, many gas detectors show an inherent fluctuation which is less than predicted by Poisson law. It simply means that when an important part of the energy of the particle is lost in the gas, there is a correlation between the number of collisions experienced by the particle and the energy lost on average in each collision. T h e F a n o factor is introduced as an empirical constant by which the variance so found has to be multiplied to give the experimental value [ 1 ] . In the other case, the ion pair fluctuations can lead to serious limitations for thin l o w pressure detectors in which there are very few primary ionization events following Poisson-like statistics. As there are large fluctuations in the energy loss per impact, measured energy losses are broader than expected by a statistical variance only [ 1 0 ] . T h e energy loss distributions are characterized by an asymmetry with an excess of large energy losses as predicted by Landau and V a v i l o v [ 1 1 - 1 2 ] . T h e relative probability of different processes induced by 31.5 M e V protons in a proportional counter are represented in Fig. 2.

300

0

10 20 30 40 50 60 PULSE HEIGHT - ARBITRARY U N I T S

70

Fig. 2. Frequency distribution of energy losses of 31.5 MeV protons traversing a proportional counter filled with 96% Ar and 4% C 0 2 . The histogram of experimental points is compared to the theoretical Landau distribution and a gaussian distribution based on ion-electron pair statistics. From Igo et al. [3]

18

Claude Stephan

Table 2. Experimental mobilities of several ions in different gases, at normal conditions [14] Gas

Ions

Mobility (cm 2 V~ 'sec"

Ar isoC 4 H 1 0 (OCH 3 ) 2 CH 2 Ar isoC 4 H 1 0 Ar CH 4 Ar 2

(OCH 3 ) 2 CH 2 + (OCH 3 ) 2 CH 2 (OCH 3 ) 2 CH 2 isoC 4 H ,+0 isoQH^O CH4+ CH4++ 2+ 2

1.51 0.55 0.26 1.56 0.61 1.87 2.26 1.72 1.09

co co

co

2. Ion drift In gas detectors an electric field is applied across the gas volume inducing motion of ions a n d electrons along the field direction, named the drift velocity w. It results in a slow m o v e m e n t linearly p r o p o r t i o n a l to the electric field Ε and inversely p r o p o r tional to the gas pressure P. Table 2 gives experimental mobilities of several ions in different gases where the mobility μ is defined by the f o r m u l a νν = μ |

3. Electron

(8)

drift

T h e m a x i m u m energy T M of electrons results f r o m a collision between the projectile and an electron f r o m the m e d i u m following two b o d y kinematics. Ejected electrons have a statistical energy distribution of the form /(A) = - L e - ^ 2π

+ c

->

(9)

ν

In this f o r m u l a according to L a n d a u [11], λ represents the normalized deviation f r o m the most p r o b a b l e energy loss. These electrons, named δ electrons at the time of nuclear emulsions, are emitted at an angle given by a free electron a p p r o x i m a t i o n cos 2 0 = Τ/Γ Μ . They will then drift in the gas volume. D u e to their small size, the collisions with molecules of the m e d i u m are m u c h less probable t h a n for ions. Therefore these electrons move m u c h faster in gas t h a n ions by a factor 1000. Typical collection times in detectors are of the order of microseconds as c o m p a r e d to milliseconds for ions. In a formulation according to Townsend [15], the drift velocity w can be written νν=-^£τ 2m

(10)

Gas Filled Detectors

19

where τ is the mean time between collisions. The electron-atom cross sections vary with energy, due to complex q u a n t u m effects between the free electron and the electron shells of gas molecules. It means that τ will depend strongly on the electric field Ε and consequently the shape of the electron energy distribution is also Ε dependent. As an example, the drift velocity w of electrons through the gas of an ionization chamber is about 5-10 6 cm/s. Multiple collisions undergone by electrons during the drift result in a diffusion in the gas. Following the kinetic theory of gases, the proportion of electrons d N / N found in the element dx, at the distance χ from the origin after a time f, is given by a Gaussian distribution-like formula

— = Ν ^jAnDt

(11) '

y

where D is the diffusion coefficient; changes in the electron energy distribution due to the presence of an electric field result in a diffusion coefficient D dependent on the electric field E. This diffusion will result in a collection on a wider surface which will affect the position resolution. A small diffusion coefficient leads to a better position resolution.

4.

Recombination

Some of the many collisions of free electrons with gas atoms during the drift may result in charge neutralization. Recombination probability depends on charge carrier density and gas pressure. This process is of importance since the original charge is lost and will not contribute to the collected ionization signal. It can even happen that the free electron is collected by a neutral atom, creating a negative ion. Collisions between ions and neutral atoms of gas are still much more likely to happen due to their very small mean free path. Charge transfer is possible either with a molecule of its own gas or with molecules with lower ionization potential. In gas mixtures, this process is very effective and rapidly removes all ions except the ones with the lower ionization potential. Negative ions may recombine with positive ions resulting in neutral atoms. At worst, these negative ions contribute to reducing the collected signal with their opposite charge.

5. The multiplication

phenomenon

When a high electric field is applied to an electrode, the primary electrons acquire an energy larger than the ionization potential of the gas atoms. Thus, they can induce secondary ionizations and the new free electrons can in turn be accelerated and produce further ionizations, generating a small avalanche. The development of such an avalanche and its quenching depends on the intensity of the electric field and on the nature of the gas filling. The gas mixture will be chosen in order to fulfill the

20

Claude Stephan

Voltoge, volts

Fig. 3. The different regions of operation of gas filled detectors. The pulse amplitude is plotted for particles depositing different amounts of energy within the gas [16]

experimental requirements: high gain, high counting rate, fast recovery, a m o n g s t others. M o r e details are given in section IV devoted to p r o p o r t i o n a l counters.

6. Different regions of operation of gas counters The collected charge is connected to the applied voltage difference between a n o d e and cathode. Figure 3 after the M o n t g o m e r y s in 1941 illustrates the different operating modes following the value of this voltage difference [16]. At very low voltage, recombination plays an i m p o r t a n t role. When the voltage is increased, full collection can be reached a n d the detector is said to be operating in the ionization c h a m b e r mode. If the electric field is f u r t h e r increased, a multiplication process begins near the a n o d e surface and gains u p to 10 4 are reached. The detector is said to be working in the p r o p o r t i o n a l regime since the collected signal is still p r o p o r t i o n a l to the deposited charge. If the voltage is increased again, the propotionality is gradually lost. Finally a saturated gain is reached in Geiger-Müller counters, or in certain conditions a new regime called the self quenching streamer (SQS) m o d e is obtained. In the next sections the different types of detectors following these various m o d e s will be described a n d recent applications will be presented.

Gas Filled Detectors

21

III. Ionization chambers Ionization c h a m b e r s have been described by m a n y a u t h o r s [ 1 7 - 1 9 ] , T h e y have been in operation for a very long time a n d in their principle there have been n o significant changes since the beginning. This type of detector has been increasingly of interest since 1975 with the development of heavy ion nuclear physics d u e to the very high stopping power of heavy ions. Consequently hardly feasible thin a n d h o m o g e n e o u s solid detectors (scintillators a n d silicon detectors mainly) can be advantageously replaced by gas c h a m b e r s which in addition are radiation insensitive. Electronic linear amplification improvements have m a d e their use competitive.

A. Principle of operation The most simple type of c h a m b e r will consist of two electrodes m a d e of two parallel plates maintaining a static electric field in between. In such c h a m b e r s it is assumed that the applied electric field is sufficient to eliminate r e c o m b i n a t i o n effects a n d t h a t negative charges are only d u e to free electrons. T h e constant electric field intensity Ε is given by the relation

(12)

where V is the voltage applied across the c h a m b e r electrodes distant by d. T h e situation is sketched in Fig. 4. W e assume that a particle has produced n0 ion pairs at a distance χ f r o m the anode. After a time i e given by the electron drift, all the electrons have reached the anode. D u r i n g this time the ions m o v e d very little b u t induced a charge on the anode. T h e signal voltage t h r o u g h a resistor denoted VR will then be

(13)

+ V

Fig. 4. Diagram of a parallel plate ionization chamber. Parameters d and χ are shown

22

Claude Stephan

VpW"

Fig. 5. Output pulse KR as a function of collection time. The fast rise time corresponds to electron collection, the broken line represents the shape of output pulses with a time constant RC « full ion collection time

here C is the ionization chamber capacitance. By waiting long enough so that all ions collect on the cathode, the maximum expected signal is

Κmax

"of C

(14)

As these collection times are of milliseconds, for high counting rates, it is interesting to choose a much shorter collection time constant so that the collected pulse reflects only the electron drift. The output pulse as a function of time is represented in Fig. 5. Unfortunately, the collected signal VK depends on the position χ of the pair creations. To get rid of this dependence, Frisch has designed ionization chambers with a screening grid.

1. The gridded ionization

chamber

The dependence of the pulse amplitude on the position of the ion pair creation can be removed with a chamber equipped with a "Frisch" grid. In this chamber the electrons have to traverse a grid set between the interaction volume and the anode. This grid is as transparent as possible, compatible with a good shielding efficiency. The grid is kept at an intermediate potential. Thus the pulse amplitudes obtained from collected electrons are not changed by the image charge of ions. An extensive study of grid effects by Buneman, Cranshaw and Harvey allows the shielding efficiency of a Frisch grid to be evaluated [20]. σ =

p+

Ρ (s/2n)(p2/4-\ogp)

where ρ =

Ina

(15)

In this formula, σ evaluates to what extent the grid will shield the collector anode from the electric field in the ionization chamber; ρ is the distance grid-anode, s is the

Gas Filled Detectors

23

wire interval, and a is the wire radius. For example, a Frisch grid made of wires 50 μπι in diameter, 1 mm apart, distant from the anode by 1 cm and from the cathode by 5 cm, will give σ = 0.971. This value is very satisfactory, since the grid has still a 9 5 % mechanical transparency. Further, collection of electrons by the grid can be substantially avoided if the electric field set between grid and anode is somewhat larger than the ionization chamber electric field. All lines of electric field by-pass the grid under the condition (16) derived by Buneman et al. [20], Κ ~ va > Ρ + PP + 2σΡΡ VG — VG d — pd — 2 σρρ

^

In this condition valuable for efficient shielding (σ should be close to 1) VA, VG, VP are, respectively, the potentials applied to anode, grid and cathode; d is the distance grid-cathode.

2. Wall and window

effects

The metallic walls of the chamber at a potential V = 0 distort the lines of force of the electric field, and consequently a fraction of the electrons are not collected. If the walls are made of isolating material the situation is still worse since their potential can change with position and time, due to electrostatic charging up. This problem has been soved a long time ago with a guard ring surrounding the collector electrode on all sides. This name comes from the days when the ionization chambers were most often cylindrical, with a radioactive source deposited on one of the electrodes. Nowadays, usually the detected particle is emitted by an irradiated target from outside a parallelepipedic chamber. The particle reaches the ionizing volume through a window made of a thin plastic film. Guard rings on the edges reduce the size of the effective volume. In front of the entrance window, they create a dead section where the energy of the particle is not measured. A clever solution consists of constituting an electric shielding with wires or aluminum strips evaporated on a plastic foil at a potential which varies linearly between the potentials of both electrodes. The efficiency of the shielding is a function of the distance between strips. A reasonable evaluation is to consider that electric field inhomogeneities disappear at a distance five times larger than the distance between strips [21].

3. Observed

signals

These detectors are mostly used in nuclear physics to detect heavy ions, to measure their total energy E, but also their energy loss AE in a small part of their total range, in order to determine their atomic number Z. As compared to solid state detectors, no pulse height defect on the heavy ion mass can be observed [22]. Indeed, the energy loss in a given medium thickness is a function of Zc with c ~ 2. For this purpose, the anode is

24

Claude Stephan

Fig. 6. Schematic drawing of a parallel plate ionization chamber, showing the splitting of the anode in 3 segments and the potential distribution on the entrance window. The voltages are given for a ionization space 5 cm wide filled with isobutane at a pressure of 0.2 atm [24]

split in two or more parts to be able to identify in Ζ and measure the energy of all the heavy ions of interest [23]. In addition the gas pressure can also be adjusted. A schematic drawing of a typical chamber is represented in Fig. 6, showing the different electrodes, the electric shielding and typical voltage values [24], Signals on the anodes due to electron collection must be amplified. The contribution of the electronic noise to the resolution is small in most cases, (10~ 4 of the collected signal typically). The actual resolution of the energy measurement is then given essentially by the fluctuations in the number of collected electrons and by energy straggling. For the total residual energy left in the detector, a resolution better than 1 % is always obtained and resolution of 0.3 % can be achieved. In the evaluation of the total energy loss in the ionization chamber one must be aware that the entrance window is not usually flat due the gas pressure inside the chamber and vacuum outside [25], As a result the residual energy left in the chamber will vary with the position height and should be taken into account. Energy loss resolutions may not be as good for heavy ions as they used to be for light particles. This fact is connected with the charge state distribution of the projectile during its path through the medium. The charge state of the particle is determined by the competition between various atomic collisions processes, essentially electron capture, ionization, which cross sections depend on the relative velocities of the projectile and the bounded electron involved [26], When velocities are similar, the projectile may undergo a few charge state changes during its trajectory inside the gas. Each time the stopping power will be modified. It means that each particle might have a different energy loss in a given gas volume. This is not true for the full range since all

G a s Filled Detectors

25

E in (MeV/u ) Fig. 7. Measured relative straggling of the energy deposition of Ar and Xe ions (dots), compared to standard collision theory (solid line) and to Badhwar theory (broken line). F r o m Pfüzner et al.

[27]

the energy of the particle is measured whatever the succession of involved processes, but is important when gases are used to identify the atomic number of the particle by energy loss determination. As a consequence of atomic processes involved, energy loss resolution in a detector will improve when the energy of the heavy ion increases but get worse with large Ζ values. Much better energy loss resolutions are obtained with relativistic heavy ions. Ions are then fully stripped most of the time so that the effect mentioned above coming from charge state changes is insignificant. Further, experimental results show a still better resolution than predicted by the collision theory presented in section II.A.2. Pfützner et al. [27] have measured energy deposition spectra using 4 0 Ar, 8 6 Kr, 1 3 6 Xe, 1 9 7 Au ions with energies between 100 and 950 MeV/u. They obtain a very narrow energy loss distribution, Gaussian in shape, in place of a Vavilov-Landau shape. The deviations from the standard collision theory can be understood by taking into account the high energy δ electrons which escape from the active volume. These energetic electrons contribute very little to the energy loss but participate to the energy straggling. Figure 7 shows the experimental results, a comparison with the standard theory and with a theory due to Badhwar taking into account the escape of δ electrons [28].

B. Various designs of ionization chambers Since the renewal of ionization chambers due to the development of heavy ion physics, many detectors have been built in large dimensions, exceeding 100 cm in depth. Most of the time they are associated with other types of detectors in order to obtain a complete

26

II PPAC (MUIPC)

0.04 mg/cmsq

III

GRS F I L L E D

DE - Ε

IONIZATION

SCINTILLATION

CHAMBER

HOOOSCOPE

>100mg/cmsq

PLASTIC

600 mg/cmsq & 12000mg/cmsq

Fig. 8. A side view of the HILI detector showing the different layers and two-dimensional plots obtained by correlations between different layers used to identify various reaction products for the reaction 7 9 B r + Al [29]

identification of all reaction products in Z, A, energy and angle of emission, as shown in Fig. 8 [29],

1. Transmission

chambers

Figure 6 represents a classical ionization chamber in which the particle flux is parallel to the electrodes. Another design consists of building chambers in which the particle traverses the detector perpendicular to the electrodes. Such chambers have been used as first detectors of telescopes in which they serve for Ζ identification only. Very thin ΔΕ detectors for heavy ion experiments can be realized using thin entrance windows and low pressure filling gases. As the particle is not stopped in the active volume of the

Gas Filled Detectors

27

Fig. 9. Backward transmission ionization chambers of the 4π detector INDRA. With courtesy of part of the I N D R A collaboration (GANIL Caen and D A P N I A - C E A Saclay)

chamber, the ion pairs are created all along the chamber crossing. T h e signal they induce on the electrodes is then not position dependent, making the use of a Frisch grid unnecessary. An example is given in Fig. 9 which represents a section of backward transmission ionization chambers of the I N D R A set-up. This 4π detector is designed to identify any ion from Ζ = 1 up to 20 and to measure their energy above 1 MeV/w threshold. It has been realized by means of three member telescopes: a transmission ionization chamber, a transmission silicon detector and an iodine cesium crystal. Dead zones on the edges have been reduced to 2 m m between telescopes. Each cell of the ionization chamber is only 5 cm wide, consequently the particle is never very far from the walls. T h e signal induced by the ion pairs is then shared between these walls and electrodes. T o avoid a few per cent error on the measured energy loss from this effect, a loose shielding grid made of thin wires every 5 mm is added in front of the anode.

2. Position sensitive ionization

chambers

The most recent ionization chambers have been designed mainly for heavy ion detection. They were built in large dimensions, either because they must cover the total focal plane of a magnetic spectrometer, or time of flight determination is needed simultaneously and then the detector is set at a distance of the order of one meter from the target. Additional requirements have been introduced: angle of entrance in the detector, perpendicular position y, the possibility of accepting more than one particle per event. Sann et al. [30] have built a conventional ionization chamber, but with a further grid half-way between the Frisch grid and the anode. This extra grid is made of wires

28

Claude Stephan

oriented following equidistant radii from the target, separated by one degree. The electrons drifting through the grid induce a charge signal on the closer wires, giving the angle of entrance. The position in the perpendicular direction is obtained by the drift time of electrons from their origin to the Frisch grid. Another way of determining the position is to equip an ionization chamber with a stripped cathode like the detector designed for the mass separator L O H E N G R I N at Grenoble [31]. In this detector, the cathode consists of an etched printed circuit board where copper strips, parallel to the beam, of 5 mm width alternate with isolators of 2 mm width. A further development consists of an internal sectorization of the detector to allow an improved counting rate capability and multiple event detection [32], MUSIC (multiple sampling ionization chamber) detectors are in use, at GSI, Darmstadt [33], associated with magnetic analysis like FRS and ALADIN.

3. Bragg chambers In order to identify heavy ions in a large range of atomic numbers and energies, it would be necessary to split the anode in as many electrodes as needed. To prevent this, Gruhn et al. [34] have proposed a new type of detector called the Bragg chamber. The stopping power of a particle varies with the energy lost in the medium and becomes maximum at the end of its range. The position of this maximum, called the Bragg peak, depends on the particle atomic number and its incident energy. Authors have built an ionization chamber like the transmission chambers described above but in which the heavy ions are stopped [34-37], The charges collected on the anode are analysed in time, allowing the determination of the Bragg peak. On the other end, the integral of all collected charges gives the total energy of the ion. This information is obtained either by digitalizing the signals, or by splitting the analogic signal into two parts with different time constants. It is important to choose a filling gas with a fast drift velocity in order to prevent recombinations and allow a reasonable counting rate. Gramegna et al. [38] have realized such a device which at 20 000 counts per second still gives AE/E = 0.9% and ΔΖ/Ζ = 50 for 5 MeV/u 32S incident beam. An alternate solution consists of using a classical chamber with a transverse electric field equipped with a resistive anode allowing the center of gravity of the Bragg peak to be determined with enough precision. A better solution has been adopted by Mittig et al. [39] for a 70 cm deep ionization chamber in which the resistive anode is replaced by 2 mm wide conducting strips connected to a delay line. A time measurement indicates at which strip the ion has stopped. The total energy is obtained by summing the signals collected on the grid and the cathode. This device has been able to separate 30 MeV/u heavy ions up to Xenon easily as can be seen in the two-dimensional plot of Fig. 10.

4. Direct current ionization chambers Ionization chambers have been widely used for radiation dose measurements, i.e. gamma-ray exposure, radiation calibration, and accelerated beam monitoring. In direct current mode, it is possible to collect either free electrons or negative charges (see

Gas Filled Detectors

29

Energy Fig. 10. Two dimensional plot of heavy ion ranges as a function of energy in Mittig et al. detector [39]. All ions up to Xe can be identified section IV.Β for m o r e details). Therefore any gas can fulfill this purpose, including those with high electron a t t a c h m e n t probability. It is then very convenient to use air at atmospheric pressure as filling gas, if a denser gas is not absolutely needed. T h e ionization current can be measured with an electrometer by sensing the voltage d r o p t h r o u g h a series resistance of at least 10 9 Ω placed between the two electrodes. F o r a m o r e stable amplification of the signal, the direct current can be first t r a n s f o r m e d into an alternative current by collecting the ion current across a R C circuit with a long time constant. T h e ionization current is also m e a s u r e d very often by integration m e t h o d : a capacitance is charged by the ionization current a n d the voltage change is measured. W h e n the voltage variation reaches a given value, the integrating system is reinitialized. This m e t h o d allows variations in time of the ionization current to be followed.

IV. Proportional counters A. Principles of operation 1. Gas

multiplication

In a p r o p o r t i o n a l counter gas multiplication occurs when a high electric field is applied. T h e principle of o p e r a t i o n of these detectors is based on the secondary ionizations

30

Claude Stephan

created in collisions between accelerated electrons and neutral gas molecules. Further ionizations will then induce new collisions, generating the avalanche. The process terminates when all free electrons have been collected on the anode. The multiplication factor Μ is usually of many thousands and is kept proportional to the energy deposited by the incident particle. A complete description of the principles of operation is given in reference [9], Gas multiplication requires large values of the electric field, of the order of 105 V/cm at normal pressure. For this reason the anode is usually a fine wire placed at the center of a cylindrical counter. In such a geometry, the electric field E(r), at distance r from the center of the wire, is given by the equation (17)

where V is the voltage applied between electrodes, a is the anode radius and b is the inner radius of the cathode. The required large values of Ε are reached for very low r values, that is in the immediate vicinity of the wire. Various authors have given analytic formulae to relate the multiplication factor Μ to the parameters of the detector [40], Diethorn gives the following expression (18) which assumes a linear relation between the electric field and the first Townsend coefficient (mean number of secondary electrons produced per length unit by a free electron) [41],

, lnM

V ln2

ΑΛ

=

iMbJä)ÄV

In

(18)

. ... , - l n Κ pa ln(b/a)

where ρ is the gas pressure, Κ and Δ Κ the potential variation between two successive ionizations, are constants for a given gas. At first approximation the gas multiplication varies exponentially with voltage V. Consequently, the applied voltage must be extremely stable to ensure a good energy resolution. For the same reason, the gas pressure must also be very stable. Table 3, a compilation due to Knoll, gives Diethorn parameters for a number of gases [42]. Table 3. Diethorn parameters for different gas mixtures Gas mixture

Κ χ 10 4 (V/cm-atm)

AV (eV)

Reference

90% Ar, 10% C H 4 95% Ar, 5% C H 4 CH 4 (methanc) C 3 H 8 (propane) 96% He, 4% C 4 H 1 0 (isobutane) 75% Ar, 15.% Xe, 1 0 % C 0 2 69.5% Ar, 19.9% Xe, 10.7% C H 4 64.5% Ar, 24.7% Xe, 10.7% C 0 2 90% Xe, 10% C H 4 95% Xe, 5% C 0 2

4.8 4.5 6.9 10.0 1.48 5.45 5.45 6.0 3.62 3.66

23.6 21.8 36.5 29.5 27.6 20.3 20.3 18.3 33.9 31.4

43 43 43 43 43 43 43 43 44 44

Gas Filled Detectors

31

Amplitude of the output pulse is proportional to the collected charge. The expected distribution in the number of electrons produced in an avalanche from a single electron is exponential in shape as given in equation (11). In strong electric fields the electron ionization probability cannot be completely independent of its past history. Finally the overall pulse amplitude distribution approaches a Gaussian shape peak [45]. This amplitude is subject to fluctuations due, like in an ionization chamber, to the number of primary electrons collected, but also to inherent fluctuations of the multiplication factor [25],

2. Detected signal The type of analysis developed in section III. A for ionization chambers can also be applied to proportional counters with several m a j o r differences. The whole process develops within a few micrometers from the anode wire. As a result multiplication will take place in a few nanoseconds. O u t p u t pulses essentially due to signals induced on the anode by the ions during their fast movement in the region of high electric field are thus obtained in a very short time, very much shorter than the total drift time. By terminating the counter with a resistor, one can get a differentiated signal allowing high counting rate capabilities.

B. Gas operation 1. Choice of gas filling As ionization and avalanche multiplication occur in all gases, the choice of gas filling could be of little importance. F o r ionization chambers it is somewhat true and the stopping power of the gas determines the choice. In proportional counters conflicting experimental requirements appear: high gain operation, low voltage, high counting rate, good proportionality. The choice will then be between different families of gases, each of them fulfill some of the requirements but not all. The family of noble gases gives avalanche multiplication at lower electric fields. Complex molecules, on the contrary, need much higher values of electric fields since there are many non-ionizing multiplication modes in polyatomic gases. Among these, most used gases are c o m p o u n d s like ammonia, methane, isobutane, carbon dioxide and tetrafluoride. The two families are characterized by a very low electron attachment probability. O n the contrary, gas molecules exist which present a tendency to attach themselves to free electrons, creating negative ions. Halogen gases as well as freons, oxygen and water vapor which have large electron affinities belong to this category. One will then think to use noble gases for gas filling. However these gases d o not ofler large gains without entering into a permanent discharge regime. Indeed the excited atoms of these will return to their ground state only by emitting energetic photons which will interact with the cathode, initiating a post avalanche due to extracted photo-electrons. Complex molecules need much higher fields but absorb the photons emitted by the excited atoms, making secondary emission very unlikely. The

32

Claude Stephan

X / p V/cm/mm. Hg Fig. 11. Drift velocity of electrons in argon-methane mixtures as a function of the reduced pressure expressed in V/cm-torr. F r o m English and H a n n a [47]

molecules dissipate the energy either by elastic collisions, or by molecule dissociation or polymerization. The solution which is chosen most of the time is to use a gas mixture, taking advantage of the low fields needed by noble gases, and of the quenching effect of complex molecules: as metastable level excitations of noble gases lie above the ionization potential of complex molecules, collisions between the two kinds of atoms produce delayed ionization. The charge is then transferred from noble gas to complex molecules. For example, the addition of methane to argon allows photo-induced effects to be suppressed by absorbing photons, with the other advantage that the electron drift velocity can be drastically increased [46], as seen in Fig. 11 [47], Difficulties can occur after some operation time when products of recombination of complex molecules are polymers. These products will deposit on cathodes and anodes creating a thin isolator layer substantially modifying the way of operation. High densities of charges develop and a permanent discharge is induced. This problem can be solved by adding molecules creating negative ions. It seems obvious to avoid gas molecules which present a tendency to attach free electrons, creating negative ions, (oxygen, water vapor, halogen gases). It means in particular that it is important to prevent the gas filling from being contaminated by air leaks with the outside. However it has been recognized for a long time that the addition of a small quantity of such

Gas Filled Detectors

33

molecules like ethylic, methylic or isopropylic alcohol has a beneficial effect on the ageing of proportional and Geiger-Müller counters [48]. The reason is that alcohol molecules have a low ionization potential and all charges will be finally transferred to these molecules which dissociate when neutralized, avoiding the formation of polymers. Analog qualities are found in a mixture k n o w n as the "magic gas" which gives exceptional ageing capabilities to proportional counters. More details on this mixture are given in section IV.C.2.

2. Gas

purification

A high degree of purity is required in all types of gas counters, since the drift time and signal heights are very sensitive to pollutions, due to effects like scattering, electron attachment and different ionization potentials of the impurities. The most important pollution comes from oxygen, nitrogen or water. O n request, quality of commercial gases is usually sufficient but in detectors, impurities come from the degassing of molecules adsorbed in chamber materials. Softeners included in many components like plastics, cables and glue are also pollutants [49], O n e must also take into account the degradation of organic filling gases under heavy irradiation. F o r all these reasons it is necessary either to ensure a continuous gas renewal, or to recycle the filling gas after purification. H o f m a n n et al. [49] suggest the use of a combination of Active C a r b o n , Hydrosorb and Oxisorb cartridges for purification. The sequence of cartridges is important.

3. Gas pressure

regulation

Gas renewal implies keeping gas pressure constant with enough precision to maintain a steady multiplication factor in proportional counters and keep energy loss constant enough for Ζ and energy determination in ionization chambers. A mechanical device exists in which a volume is enclosed at the given reference pressure. A limber wall of the volume is in contact with the detector volume. A variation in the detector pressure induces deformation of this wall on which a needle is fixed. The needle opens or closes a hole through which gas can enter into the detector to reduce the pressure difference. A more flexible system consists of comparing the detector pressure to a reference pressure (or vacuum) with a sensor. An electronic design allows progressive valves to be opened either to introduce more gas or to p u m p the filling gas as shown in Fig. 12. A detector can then be maintained at the chosen pressure with a precision of 1(T 3 .

C. Various types of proportional counters Proportional detectors give high amplitude signals due to the multiplication factor. For this reason they were first used for energy loss measurements. However, because of

34

Claude Stephan

VG X JL

VV Detector

Iί

Refer, pressure

Sensor

Ϊ Regulation

t

Vacuum

Gas

Fig. 12. Principle of operation of an electronic gas pressure regulation

variations in the gas amplification, the energy resolution is seldom better t h a n a few per cent (20% for protons), a n d p r o p o r t i o n a l detectors are not competitive with m o r e recent types of detectors. T h e y are m u c h m o r e effective for timing and position determination [50],

1. Position sensitive proportional

counters

P r o p o r t i o n a l counters have been widely used t o replace nuclear emulsions as active focal plane detectors of magnetic spectrographs. The position resolution is then the most i m p o r t a n t characteristic. W h e n an incident particle hits a c o m m o n cylindrical p r o p o r t i o n a l counter, the electrons drift f r o m the place of f o r m a t i o n to the a n o d e wire following radial field lines. As the avalanche develops in a very small zone of the wire, the position of the cascade indicates the longitudinal position of the passage of the particle t h r o u g h the detector. A position sensitive p r o p o r t i o n a l detector can be considered a distributed RC line. A m e t h o d to get the position on the wire is to observe the relative rise time f r o m signals out of preamplifiers placed at the extremity of the a n o d e wire. Excellent spatial resolutions have been obtained with voltage sensitive preamplifiers by this m e t h o d introduced by Borkovski and K o p p [51]. However, with this technique, end effects restrict the useful length of the counter. T h e most c o m m o n m e t h o d of position sensing is the charge division method. T h e a n o d e is m a d e f r o m high resistive wire so that the collected charge is split between the preamplifiers placed at either end, with an amplitude p r o p o r t i o n a l to the relative position of the avalanche. Different species of resistive wires typically of 20 μηι in diameter and of a few h u n d r e d Ω cm ~ 1 resistivity are used: stainless steel, q u a r t z fiber coated with carbon, or nichrome. The differential linearity will rely on the homogeneity of the wire. T h e position x, f r o m one edge of the detector, is then calculated by the

Gas Filled Detectors

35

relation (19) where Qlt Q2 are the collected charges on each side of the wire and L its total length. Proper adjustment of gains are needed for optimum resolution. Various effects contribute to the finite position resolution: electronic noise, thermic noise of the wire, multiple scattering, energy loss fluctuations. Charge preamplifiers used to collect charges do not need to have a good resolution since the electronic noise is negligible compared to the thermic noise of the resistive wire. The feed-back resistor ρ of Fig. 13a can then been reduced by a factor 10 to 20 ΜΩ in order to eliminate low frequency noises at the entrance of the preamplifier. Nevertheless they must have a good linearity, better than 0.1%. In these working conditions, the voltage variation depends not only on the current collected on the wire but also on the equilibration of capacitive charges through the wire. If for example the wire resistor is R = 4 kQ, the time constant of this contribution is RC2/2 = 20 μί, low enough to draw a variation of the position determination with the rise time of the pulse. Consequently it is more convenient to work with the design of Fig. 13b. In this configuration the voltage variation at the exit of the preamplifiers is the product of a function of exponent

Out In

(a)

Out

Fig. 13. (a) standard functional diagram of the first stage of a charge preamplifier, (b) actual functional diagram for the particular use of charge division.

36

Claude Stephan

2 p C 1 = 320 μχ by a sinusoidal function of period 2π

Τ = 1 pC1RC2

580 μί

(20)

1 4 (pCja

T h e F W H M c o n t r i b u t i o n to the limitation in position resolution d u e to thermic noise is [52],

(21)

where C is the capacitance of the detector, C L is the load capacitance, Q is the total charge generated in the detector, k is Boltzmann's constant, Τ is the absolute t e m p e r a t u r e a n d C s = C + 2C L . E q u a t i o n (21) shows that the thermic noise contribution decreases for larger collected charges. T h e entrance w i n d o w gives the main multiple scattering contribution. The effect is emphasized for particles incident u p o n the counter at a large angle, as usually the case in magnetic spectrometers. As the position is determined by the centroid of the charge distribution deposited by the particle along its trajectory, energy loss fluctuations can also degrade the resolution for large angles. T h e c o m b i n a t i o n of all contributions allows resolutions of p r o p o r t i o n a l counters of a few tenths of millimeters to be expected [52], A n u m b e r of magnetic spectrometers have been equipped with such devices [ 2 4 - 2 5 , 5 3 - 5 6 ] ,

2. Multixvire proportional

counters

A p r o p o r t i o n a l c o u n t e r is useful to detect with a fast response a particle in a limited region of space. T h e idea to stack a n u m b e r of such detectors is n o t very attractive so the possibility of putting multiwire structures in the same volume of gas was examined. It was t h o u g h t t h a t such a design could not work properly since the signal obtained on the wire where the avalanche occurred would spread, by capacitive coupling, into all wires. It was the merit of C h a r p a k et al.51 to d e m o n s t r a t e that the positive induced signals in all neighbouring wires largely c o m p e n s a t e the negative signals due t o capacitive coupling. A multiwire p r o p o r t i o n a l c o u n t e r ( M W P C ) consists of equally distant thin wires sandwiched between two c a t h o d e planes. Each wire can have its own electronic channel. A schematic design is sketched in Fig. 14. Immediately after their discovery, the r e m a r k a b l e properties of these detectors have stimulated systematic studies. Their properties, that is efficiency, time resolution, position resolution, as a function of a n u m b e r of parameters, mechanical dimensions, wire diameter, high voltage, gas mixture have been extensively investigated [58]. T h e properties of such c h a m b e r s depend strongly on the choice of geometrical parameters. T h e electrical characteristics, that is electric field variations, capacitance, are determined by the spacing s between a n o d e wires, their radius a and the distance

Gas Filled Detectors

37

cathode

Ϊd ·

!

·

·

·

·

·

·

·

a n o d e wires

cathode Fig. 14. Schematic diagram of a multiwire proportional counter. A set of parallel wires is mounted symmetrically between two cathode planes. Geometrical parameters s, d are shown

d from the wires to the cathodes. The equation (22) after Erskine gives the capacitance per unit length [59] C=

^ (nd/s) — In (2 πα/s)

(22)

This equation (22) shows that the collected charge CV0 decreases when the wire spacing s decreases. The voltage applied V0 must be increased to obtain the same gain. For example if sis changed from 2 mm to 1mm, K 0 must be multiplied by 1.5 at least, which means more difficulties of operation of the detector. The solution can lie in reducing the wire diameter, but there are obviously mechanical and electrostatic limitations. Practically, good operating conditions are obtained in a 8 mm gap, with wires between 10 and 20 μηι in diameter and a 2 mm spacing. M W P C have been used from the beginning as position sensitive detectors. The precision was given by the distance between wires. The position in the y direction was obtained with another chamber in which the anode wires were perpendicular to those in the first chamber. A better two-dimensional position precision can be achieved by measuring the charge distribution induced by the anode avalanche on the cathode plane. Indeed, a large part of the negative signal collected on the anode is not due to electrons but to the drift of positive ions towards the cathode. This movement induces positive signals on all adjacent electrodes. Thus, the gravity center method gives extremely good localization resolutions. A spatial resolution of 60 μηι has been achieved with 400 μηι spaced wires [60], By making cathodes with metallic strips or two perpendicular wire planes, it is possible to obtain the two coordinates of the avalanche in one MWPC. The resolution in both directions is not equivalent, however much better than the wire spacing because, under moderate gain conditions, the avalanche remains localized at the point where primary electrons reach the wire [61], and does not surround the anode symmetrically. Therefore, the center of gravity of the induced charge distribution contains information about the position of the original

38

Claude Stephan

ARGON - C 0 2 100

beginning yW of y A ploteau j||||||

0« ftl Ο Ο

limit / of proportionality —

-

50

Λlimit

; jJl/ 1

0 4Ö00

1 6000

HIGH

of

breakdown

-

I volts VOLTAGE

Fig. 15. High voltage plateau and breakdown voltage as a function of C 0 2 and isobutane concentration in argon mixtures. From Bouclier et al. [58]

ionization. Further, the ions due to multiplication follow the same electric field lines which terminate where the ionization started [62], Multiplication factor and consequently efficiency depend on the voltage and on the filling gas. Systematic studies of operating parameters for different mixtures at various gas concentrations have been carried out. Position and length of the plateau of ^ 9 9 % efficiency up to the end of the proportional region, break down voltage, were determined. Studies by Bouclier et al. [58] are represented in Fig. 15 for argon-isobutane and a r g o n - C 0 2 mixtures. Gains up to 106 can be reached with a r g o n - C 0 2 mixtures. These authors have found that addition of electronegative gas Freon gives remarkable properties to M W P C chambers. With a special mixture named "magic gas" by the authors: argon, 70%, isobutane, 29.54%, 0.46% freon-13Bl (CF 3 Br) there are important changes in the behaviour of the detector: gains increase, heavy irradiations induce no change in the characteristics, however no spatial improvements are observed, but time resolution is somewhat degraded.

Gas Filled Detectors

39

These detectors were first developed for particle physics experiments, but they have been widely utilized in nuclear physics as position sensitive detectors [63], many of them in focal planes of spectrometers [64-66], A cylindrical M W P C with a large solid angle has been built for fission fragment angular distribution studies [67], They are also very useful for low energy X-ray detection with position readout, allowing quite high counting rates [68-69]. G. Giorginis et al. [70] have built a pressurized Helium M W P C as a neutron Polarimeter, using the scattering reaction (n, 4 He), the Helium gas serving as α target to the neutrons as well as for detection of the He-recoil track. They are also widely in use for beam profile measurements of accelerated beams.

3. Detectors operating in the self quenching streamer mode Systematic studies of proportional chambers led, under certain conditions, to the observation of abnormally large signals. These pulses could not be attributed to a Geiger-Müller mode because their duration, a few tenths of a nanosecond, was too short. They were obtained with the argon-isobutane mixture, but with large anode wires. The signals observed had the same characteristics as those found with the "magic gas" described above. Alekseev et al. [71] have shown that both effects belong to a definite operating mode above the proportional region, different from the GeigerMüller mode, and called the self-quenching streamer mode (SQS). In this process, due the presence of a strong electric field, the primary avalanche around the wire develops in a streamer perpendicular to the anode, directed towards the cathode, following the trajectory of primary electrons. The discharge develops till 1 to 3 mm from the anode. It does not result in a spark breakdown because of the quenching effect of the gas which becomes efficient as soon as the weak electric field region is reached. Dimensions of the streamer imply a large number of charges, which explains the difference in pulse height as compared to the proportional mode. The main features are: a very stable operating mode due to a wide efficiency plateau of more than 1000 V, pulses shorter than 200 ns, with a 20 ns rise time, high pulse amplitudes but with a large dispersion, giving a very good signal to noise ratio. Signals depend very little on the primary ionization. Lower mechanical requirements on the geometry make the design of large chambers easier. These interesting characteristic have been widely exploited in particle physics (streamer tubes), but very little in nuclear physics detectors. However, multiwire proportional chambers working in the SQS mode have been built for the focal plane of the spectrometer SPES1 at the SATURNE facility [72], allowing detection of different ions with the same operating conditions, with a spatial resolution of 0.4 mm over a one meter length.

4. Drift chambers Multiwire proportional counters achieve excellent position resolution. However in heavy ion physics many experimentalists prefer to use drift chambers which have several attractive features: a relatively simple and inexpensive design, a reduced mass density, a better differential linearity, the ability to accept a large variety of ion species,

40

Claude Stephan

precision of the drift distance measurement; some structures display no wires on the ion trajectory. In the simplest design, a drift chamber is constituted of two different parts, as shown in Fig. 16: a drift region working in the ionization chamber regime and anode wires where avalanche multiplication occurs. Α Frisch grid can separate the two regions. Ion-electron pairs created by the particle in the uniform electric field region migrate towards the proportional detection region. The two dimensional localization is given by avalanches on anode wires and by the drift time. This last one is measured between a prompt timing signal derived from an additional timing signal and a pulse collected on one of the electrodes. The avalanche localization is obtained with a proportional resistive wire [73]. In the solution adopted for the spectrometer S P E G at G A N I L , signals induced by the avalanche on the proportional counter wire are collected on a stripped cathode [74]. Each element of a delay line similar to those described in section V.A. is connected to a strip, and position resolution of 0.4 mm has been achieved. However, for large detector surfaces these designs would lead to uncomfortable working voltages and too long drift times. Drift chamber structures derived from M W P C consist of a wire plane made of alternating anode and cathode wires corresponding to a number of cells [75]. A delay line connected to the anode wires determines the avalanche anode wire position. This set-up has been adopted for some spectrometer detection systems [ 7 6 - 7 8 ] . In the vertical drift chamber of the M I T energy-loss spectrometer, electrons drift from a high voltage cathode plane towards the sense wire. 120 μιτι position resolution has been obtained, achieving the Ap/p < 10 4 requirements [79]. Alternate cathode signals can be bussed out on Odd and Even outputs and coded separately [78], These arrange-

Gas Filled Detectors

41

ments allow left-right ambiguities to be resolved, i.e. distinguishing between t w o tracks located at the same distance from, but on either side of, the a n o d e wire. T h e y allow the incident angle to be reconstructed [80]. In a n o t h e r design, in order to get a c o n s t a n t electric field along most of the electron drift trajectory, a n d thus a c o n s t a n t electron velocity, a n o d e wires have been centered in a symmetric cell limited by c a t h o d e wires at a decreasing potential on each side of the a n o d e [81], All these detectors are a b o u t one meter in length.

5. Pictorial drift

chambers

In m a n y nuclear physic fields of research, inclusive m e a s u r e m e n t s have p r o v e n to be insensitive to underlying reaction mechanisms. Exclusive measurements have b e c o m e necessary for a better understanding and have led to the design of 4π detectors, also useful for rare event detection. A solution was f o u n d in building streamer chambers to obtain high multiplicity reaction details. Within one microsecond after the particle has passed t h r o u g h the detector, a very high voltage accelerates primary electrons a n d visible tracks f r o m streamers are obtained. These tracks can be p h o t o g r a p h e d for subsequent analysis. Streamer chambers are limited to a few counts per second. Review papers are given in references [ 8 2 - 8 3 ] , A m o n g all 4π detectors designed for high interaction rates, pictorial drift chambers represent p r o b a b l y the most sophisticated gas detectors ever built for nuclear physic experiments.

U22 cm

Ε II

Fig. 17. Drawing of one sector from the PDC D I O G E N E in use at SATURNE [84]

42

Claude Stephan

A pictorial drift chamber consists of a large gas volume of the order of one cubic meter, usually cylindrical in shape, in which ionization electron drift, under the action of a homogeneous electric field, towards a radial plane of multiplying wires parallel to the beam axis. Like in drift chambers described in section IV.C.3, one obtains the position of the particle in one direction by the wire number, and in the other direction, by the drift time. The third coordinate, along the wire, can be deduced by charge division, as described in section IV.C.l. The energy loss is also determined. Particle identification is obtained by means of energy loss measurement versus momentum analysis from Bp determination. The large drift time can be inconvenient for high interaction rates, so the detector is divided into several identical sections. Figure 17 shows one of the ten radial sectors of the P D C D I O G E N E is use at S A T U R N E [84], Each sector comprises 16 drift cells. A one tesla magnetic field, parallel to the beam axis, allows the magnetic rigidity of the particle to be determined. Identification of mass and charge is obtained from the correlation between energy loss and magnetic rigidity. Some other detectors of this type devoted to nuclear physics have been in operation at T R I U M F [85], and the BEVALAC [86], A new An detector has been built to be used with the accelerator SIS at GSI: two parts of this detector are pictorial drift chambers. A Central Drift Chamber (CDC) with conical front ends, covering 30° 160° consists of 16 radial segments [87]; the H E L I T R O N covering 7.5° ^ θ < 30° consists of 24 radial drift chambers [88],

V. Low pressure gas detectors A. Parallel plate avalanche chambers Parallel plate avalanche counters (PPAC) have been known for many years to be precise timing instruments [89], but were scarcely used before the considerable development of heavy ion physics, for which gas detectors are better suited. A P P A C consists of two thin parallel stretched foils with a very low gas pressure in between. Particles traverse the detector perpendicular to the planes. The principle of operation is the same as with M W P C . The gap between the foils must be small, of the order of a few millimeters, in order to maintain a high electric field, reduce the time spread and get a good time resolution. It also has to be uniform to ensure the same operating regime on the whole active surface of the detector. Copper coated epoxy resin for printed circuit board, on which thin metallized (with aluminum or gold) plastic foils are glued, is very well suited for this purpose. The copper layer is used for electric connections. These detectors are built to operate with pressures ranging from 1 to 20 millibars. Under these low pressure conditions, a voltage of a few hundred volts applied between the plates, typically 300 V/cm-mbar, is sufficient to ensure the proportional regime. Electrons released gain enough energy to induce immediate secondary ionization in the homogeneous electric field, and a Townsend avalanche is formed [15]. It seems that pure hydrocarbons are best suited for these detectors, the highest gains being reached with isobutane [90], Multiplication factors of a few thousands are obtained. 100%

Gas Filled Detectors

43

detection efficiency can be achieved in a wide domain of energy losses. The energy resolution AE is limited to about 20% due to straggling in the gas. For large energy losses a pulse height saturation can occur. A 2 to 3 ns rise time pulse is collected due to the high velocity of electrons and the good homogeneity of the electric field. Positive ions contribute very little because they are not very close to the anode, in comparison to what happens in a wire counter. Only the fast component of the signal due to the motion of electrons is used. The slow part from the positive ions is eliminated by differentiation of the signal. Best timing performance may not necessarily correspond to the fastest output rise time, since at very low gas pressure, pulse fluctuations can become important [91]. Time resolutions better than 200 ps (fwhm) have been obtained with such detectors even in large dimensions [92], Of course a delay of 40 to 50 ps/cm from the propagation time of signals along the electrodes has to be taken into account [91]. It is possible to achieve a very good spatial resolution with avalanche parallel plate detectors. The localization is realized by strips on the cathodes and by using fast delay lines to reconstruct the center of gravity of induced charges, as illustrated by Fig. 18 [93]. For example, Fig. 19 shows a graph where a resolution better than 300 μηι has been obtained with a 2 mm wide strip detector operating at 10 mbar isobutane pressure irradiated by 84 Mev oxygen ions [94]. Position resolution depends on electronics of course, but also on the quality of the delay line. This is true not only for PPAC but also for drift chambers already described. Most authors use manufactured delay lines. Some

evaporated strips Y stop induced

charg start "Υ-cathcd·

avalanches

induced

X stop

charge

Fig. 18. Schematic diagram by Breskin and Zwang of localization by strips on cathodes in a PPAC. Fast delay lines allow to reconstruct the center of gravity of induced charges [93]

44

Claude Stephan

Fig. 19. Pulse height distribution of 84 MeV oxygen ions through 200 μιη slits in a PPAC equipped with strips 2 mm wide. With courtesy of Gaiardo et al. [94]

Fig. 20. Drawing of a delay line. Capacitors Cs take into account the detector capacitance

authors have obtained better results with electromagnetic delay lines they have built [94]. The delay line made by the authors from Fig. 19 is composed of simple LCp — Cs circuits as shown in Fig. 20. The conception of inductive elements L is important to get a low dispersion delay line [95], They are made of a continuous winding on a threaded plastic rod having taps every 3 to 5 windings to constitute individual cells. Impedances of the order of 100 Ω are obtained with a few nanosecond delay per element. The capacitor Cs eliminates the overshoot of the pulse collected at the end of the line, due to mutual inductance between cells. This capacitor Cs is detector dependent because the total capacitance takes into account the capacitance of the detector. These delay lines

Gas Filled Detectors

45

have a better rise time, a higher impedance and less amplitude attenuation than commercial ones. For example, an attenuation factor less than 2 is obtained in a detector equipped with a 75 cm line of 400 ns total delay and 15 ns rise time [56], A careful construction of the induction elements and a selection of capacitors give a good differential linearity. Nonlinearities smaller than 1 % have been reached for a drift chamber in the focal plane of the spectrometer SPEG at GANIL [74], PPACs are extensively used as first detectors in many experimental equipments to obtain a timing signal. They possess a number of attractive features: these detectors are not sensitive to radiation damage, no sparks come from wires, the fast removal of positive ions gives them a high rate capability, as position detectors they have the advantage over MWPCs that wires are absent on the ion trajectories. Their low operation gas pressure allows the use of very thin windows, which renders PPACs the thinnest timing and position sensitive detectors available. These properties have found specific applications in heavy ion experiments, essentially for ray tracing [96-00]. As they can be built in large dimensions, not influenced by strong magnetic fields, they are used very often in the focal plane of spectrometers [24,32,56],

B. Low pressure multiwire proportional counters It has been shown that a very good time resolution is achieved in very low pressure proportional counters, due to the fast drift velocity of the electrons. Breskin has demonstrated that at pressures of the order of 1 mbar or less, multiplication happens in two steps [101]. A first amplification takes place in the almost constant field region, like in parallel plate avalanche counters described in the preceding section. A second amplification happens, like in a high pressure Μ WPC, around the wire where values of 105 Volts/cm-torr are reached. In this large electric field ions recoil with velocities larger than 106 cm/s. As a result, the time distribution of the ion induced signal is divided into two parts: a very fast one mixed with the electronic signal and a slower one due to the ion drift in the constant electric field region [102]. The variation of the time distribution signal with gas pressure is illustrated in Fig. 21. At 3 torr, the pulse time distribution is symmetrical in shape due to the very fast collection of ions on the cathode. At 25 torr the tail of the pulse is due to the ion drift induced signal. Consequently, the best time resolution is obtained at low gas pressure, where the rise time is shorter. At a 2 torr gas pressure, a very good time resolution of 100 ps has been achieved by Breskin et al [103], Excellent spatial resolutions can be found; a 80 μιτι F W H M resolution has been reached with a counter in which electrodes were wound with a space of 2 mm [104].

C. Low pressure multi-step detectors Multi-step detectors are used for imaging in many fields of research or in applied physics [105-107]. The mechanism of these detectors is based on a preamplification of the initial charge and the transfer of the primary avalanche to a secondary amplification step. They are used in nuclear physics for the detection of low energy heavy ions.

46

Claude Stephan

20 · time resolution ( ns )

IS

10

5

10 ns

„

0

20 ns

11 fk 10

30 40 ρ ( mm Hg )

Fig. 21. Variation of the time distribution signals in a MWPC [102] filled with heptane at different pressures. Represented signal shapes correspond to 3 torr and 25 torr, respectively. From Binon et al.

Nuclei of ,4 ^ 100 and 10 M e V total kinetic energy are produced in fusion reactions and studied at energies close to the coulomb barrier. Indeed, a window thickness of 1 μπι is required, which limits the gas pressure of gaseous counters to 1 or 2 mbar. At the energies considered here, electronic energy loss is low and only a few tens of electrons will be created in the low pressure detector. Even P P A C or M W P C are excluded because of insufficient amplification gain. Breskin et al. [ 1 0 8 ] have studied a multi-step detector able to satisfy this particular need. The preamplification step operates in the P P A C mode. W h e n used at normal pressure, a transfer region to the second amplification zone is essential for the absorption of the photons produced in the gas mixture. At low gas pressure, electron diffusion leads to wide avalanches, leading to an efficient transfer process. Further, the quenching efficiency of isobutane renders optional the transfer region which may degrade the g o o d timing properties of the detector. In the Breskin device the second amplification stage is a M W P C . This detector gives g o o d timing resolution (150 ps at F W H M ) , 200 μπι position resolution, and was fully efficient for the detection of 1 6 0 G d nuclei with kinetic energy d o w n to 1.3 MeV.

VI. Gas microstrip detectors A. Description A microstrip gas chamber (MSGC) is a proportional counter in principle. This device could be built thanks to the development of microelectronic technology. In some way it

Gas Filled Detectors (a)

47

Drift cathode plane

(b)

Fig. 22. (a) schematic cross section of a microstrip gas detector designed by Angelini et al. [109] (b) Anode-cathode structure of the detector. The substrate is represented in black.

reproduces the field structure of multiwire chambers, but on a m u c h smaller scale a n d interesting specific designs. It consists of a sequence of alternating very thin metal anodes of a few μηι width a n d c a t h o d e strips, etched with high accuracy on an isolating substrate, as represented in Fig. 22 [109], W i t h this technology, each cell can be m a d e 200 μηι wide, or even less. These detectors are m a d e using electron b e a m lithography for mask scribing, as well as p h o t o l i t h o g r a p h y a n d thin film deposition to engrave the electrodes. A deposit of metal on the b a c k p l a n e defines a back cathode. T h e e n t r a n c e window in front can also serve as a cathode, to t r a n s p o r t primary electrons generated by ionization t o w a r d s the thin a n o d e where multiplication occurs. T h e first detector of this type was built by O e d w h o i n t r o d u c e d the idea a n d d e m o n s t r a t e d its qualities as a low energy a l p h a and p r o t o n detector [110], T h e m o d e of operation of this device is as follows: a drift electrode defines a zone of c h a r g e collection of a few millimeter thickness, then field lines connecting the drift and back c a t h o d e to the anode, concentrate on the thin a n o d e strips at a positive potential with respect to the b r o a d e r c a t h o d e strips, resulting in a high electric field of 50 k V / c m or

48

Claude Stephan

more in their n e i g h b o u r h o o d . D u e to the small width of the a n o d e and of the a n o d e - c a t h o d e separation, operation at only a few h u n d r e d volts results in the p r o p o r t i o n a l regime m o d e with gains of thousands. T h e ions p r o d u c e d a r o u n d the a n o d e are c a p t u r e d rapidly by the c a t h o d e strip at a distance of only 50 μηι. M a n y laboratories have started to study properties of these detectors in detail as a function of the gas filling, the n a t u r e of the substrate a n d of the electrodes. G a s mixtures for M W P C can also be used for those detectors operating at n o r m a l pressure. F o r improved efficiency it is suitable to fill the device with dense gases in which the energy loss is increased, to c o m p e n s a t e for small drift gaps necessary to achieve fast drift a n d minimize parallax error for localization accuracy. P u r e isobutane, argon a n d xenon mixed with quenching gases are most suitable. After irradiation, a gain modification very localized to the region of irradiation has been attributed to the well k n o w n ageing of gaseous counters, t h a t is, deposits on a n o d e strips. This p h e n o m e n o n has been observed in all mixtures by Bouclier et al. [111] except for the mixture 9 3 % argon, 7 % D M E (dimethyl ether) for which no visible deposit could be found.

B. Influence of the substrate and nature of electrodes T h e substrate should be a good enough isolator to secure a correct potential distribution. At the same time, its conductivity has to be large enough to evacuate charges arising f r o m ions formed in avalanches near the anodes. Barasch et al. [112] estimate 10 1 3 Ω/cm 2 as an optimal value for particle fluxes of 10 5 c m ~ 2 s ~ 1 a n d a gas gain of a b o u t 10 4 . It is not obvious which material corresponds to this requirement. Glasses or plastics exhibit m u c h higher resistivities, a n d when they are loaded with conductive materials, it is difficult to obtain the required resistivity. Generally, one finds an i n h o m o g e n e o u s mixture of microscopic conducting a n d dielectric phases. Glass s u p p o r t s were first used. However, after a few hours of operation, a d r o p in gain is observed, independent of the irradiation rate, accompanied by an energy resolution d e g r a d a t i o n [111]. W h e n switching off the high voltage, the gain is restored to its initial value after a few hours. This effect m a y be interpreted by a variation of ionic conductivity in glass d u e to the m o t i o n of alkali ions. In the presence of potential, as ions migrate in the bulk material, the n u m b e r of current carriers decreases with time, resulting in an increasing resistivity. This effect can be reduced by a repulsion of ions with the rear potential. In fact this m e t h o d is effective only for thin substrates. A solution can be f o u n d in using electronic conductivity glasses with a low p r o p o r t i o n of alkali like M u r a n o glass or various C o r n i n g glass samples selected by some authors. O n the other h a n d plastics like K a p t o n , Kevlar [111], and p o l y u r e t h a n e copolymer [112], seem to be suitable s u p p o r t s for M S G C , being mechanically stable, with good surface quality a n d the required bulk resistivity. A way to solve surface charging up effects is to m a k e the surface become somewhat conductive by ion i m p l a n t a t i o n or by thin film deposit. F o r example, adequate stability has been obtained on highly isolating q u a r t z by implantation of 5 · 10 1 6 Boron ions/cm 2 [ 113]. A quite different m e t h o d employed by Angelini et al. [ 114] consisted of using an a m o r p h o u s low resistivity silicon wafer as substrate. T h e isolation between a n o d e a n d

Gas Filled Detectors

glass substratum

49

metallic layer

Fig. 23. Structure of electrodes designed by Oed et al. [117], allowing a 105 gas amplification

cathode is provided by a 2 μιτι thick thermal oxide layer. This technique seems very promising [115], a n d multi-chip technology m a y well be the next step in the development [116], Devices with such thin dielectric substrates have interesting features as described in section VI.C. Using very thin anodes, the potential difference is s o m e w h a t limited because of sparks at the edges p r o v o k e d by the high field strength. As the heat capacity of a few μπι strip is limited, the spark can melt or even e v a p o r a t e the anode. F o r this reason, the layer thickness should be of at least 1 μπι, with the additional a d v a n t a g e of providing a low strip resistance, which is essential for a short rise time of the signal. Skillfully chosen structure shapes like those represented in Fig. 23 m a k e it possible to keep the field strength in the parallel sections higher t h a n in any other p a r t of the strips [ 117]. T o avoid uncontrollable charging on the isolating part, at least 5 0 % of the substrate surface should be covered with metal. Deposits of a l u m i n u m , c h r o m i u m , a n d gold are used. The former two materials form protective oxide layers which can eventually lead to b r e a k d o w n effects.

C. Performances This detector type is very promising and should enable the development of new types of experiments in nuclear physics. T h e high precision geometry of anodes, m a d e with 0.1 μπι definition by microelectronic techniques, results in a uniform gas amplification o n the entire surface. Therefore

50

Claude Stephan

ι 1.2.

μ

ι ι ι ι ι ι ι ι ι ι ι ι ι ι ι ι ι ι ι ι ι ι

_ Β 0.16 keV

μ 1.0 » c

g 0.8

o

0.6 0.4

0.2

0 0.0

0.5

1.0

1.5

2.0

2.5

Energy ( k e V )

Fig. 24. Pulse height distribution of very low energy X-rays obtained with a microstrip gas detector [118]

one can expect a much better energy resolution than obtained with wire proportional counters. However, the amplitude of the signal is limited by an avalanche amplification gain not likely to exceed 104, mainly because of spark breakdown. The compact configuration of the device and eventually the high dielectric constant of the substrate result in a high capacitive load, which causes a high noise level in the charge sensitive preamplifiers. Thus, energy resolution can be limited by the signal-to-noise ratio. The resolution is about 12% at 5.9 KeV [109], Figure 24 shows a low energy X-ray spectrum from a detector for the SOD ART telescope [118]. The most interesting property of these detectors is their good granularity due to very small spacing, ensuring a high accuracy in event localization. Systematic errors originating from the position of strips can be neglected. Various authors have found excellent resolutions σ = 30μπι for tracks perpendicular to the surface [119]. This result is obtained by calculating the center of gravity of signals given by primary electrons spreading by diffusion over two or three strips. Induced signals due to collection of positive ions on the back electrode can be exploited for two-dimensional read-out [ 120]. The signal amplitude depends on the thickness of dielectric substrate. It is then advantageous to use a design like the one employed by Angelini et al. [114] in which the isolation is provided by a 2 μηι thermal oxide layer only. However, the large distance between anode and backplane causes a very small signal on the back electrode, electrically shielded by the much closer cathodes. In new devices this distance has been reduced to the thin silicon oxide layer of 2 μηι [121], so a large fraction of the charge delivered in the avalanche is induced on the back electrode almost free of shape. Figure 25 represents the electric field lines as with a thin substrate. The back electrode can then be split in as many strips as desired to make a true high resolution twodimensional device, or even divided into square pads separated by a few μηι forming

Gas Filled Detectors

51

Fig. 25. Representation of electric field lines in a MSGC with a thin substrate inducing signals on the back cathode [115]

a pixel structure [122]. A spatial resolution of 30 μιη for both coordinates is expected in the future. The fast ion collection allows for a high irradiation flux. Rates of the order of 105 particles/mm 2 s have been measured [111, 113-114,122], with a small decrease of the amplification gain for some devices. As we have seen, these counters present many advantages as compared to the existing detectors and if they can be built in large dimensions [123], they will find many applications in the future: spatial geometry ten times smaller than MWPC, which means a position resolution improved by an order of magnitude, low voltage of operation, fast collection, high counting rate, high radiation resistance, low cost, mechanical stability.

VII. Conclusion In the 1970's, the combination of electronic improvements, the discovery of multiwire proportional detectors, and the huge development of heavy ion physics have put gas detectors among the best suited types of detectors in nuclear physic experiments. Developments of various kinds of very competitive detectors appeared during the next ten years: energy, position and timing detectors, sectorized ionization chambers.... At the same time, physical processes in gas detectors have been extensively studied and are now well understood. Studies on gas filling have resulted in the discovery of interesting

52

Claude Stephan

mixtures increasing the detector's capabilities. Nowadays, gas detectors are mostly used in combination with solid type detectors in complex designs allowing complete identification of reaction products, sometimes in 4π equipments. N e w devices are continuously appearing in publications. For example, the Proceedings of the Sixth International Wire Chamber Conference held in Vienna in 1992, contain about one hundred communications [124]. In particular, microstrip gas detectors will certainly be subjected to important developments in the near future.

References [1] Fano, U., Penetration of protons, alpha particles, and mesons, Ann. Rev. Nucl. Sei., 13,1, 1963. [2] Ahlen, S. P., Theoretical and experimental aspects of the energy loss of relativistic heavily ionizing particles, Rev. Mod. Phys. 52, 121, 1980, and references therein. [3] Northcliffe, L. C. and Schilling, R. F., Range and stopping power tables for heavy ions, Nucl. data tables A7, 233, 1970. [4] Ziegler, J. F., Handbook of stopping cross sections for energetic ions in all elements, Pergamon, New York, 1980. [5] Hubert, F., Bimbot, R., and Gauvin, H., Range and stopping power tables for 2.5500MeV/u heavy ions in solids, At. Data and Nucl. Data Tables 46, 1, 1990 [6] Geissei, H., Laichter, Y., Schneider, W. F. W., and Armbruster, P., Energy loss and energy loss straggling of fast heavy ions in matter, Nucl. Instr. and Meth. 194, 21, 1982. [7] Herault, J.,Bimbot, R., Gauvin, H., Kubica, B., Anne, R., Bastin, G., and Hubert, F., Stopping powers of gases for heavy ions (O, Ar, Kr, Xe) at intermediate energy. Vanishing of the gas solid effect, Nucl. Instr. and Meth. B61, 156,1991. [8] Klein, O., and Nishina, Υ., Z. Physik 52, 853, 1929. [9] Sauli, F., Principles of operation of multiwire proportional and drift chambers, CERN/77-09, Geneva, 1977. [10] Ermilova, V. C., Kotenko, L. P., and Merzon, G. I., Fluctuations and the most probable value of relativistic charged particle energy loss in thin gas layers, Nucl. Instr. and Meth. 145, 555,1977. [11] Landau, L. M. and Lifshitz, Ε. M., Electrodynamics of continuous media, AddisonWesley, Reading, Mass. USA, 1960, chap. 12. [12] Vavilov, P. V., Ionization losses of high energy heavy particles, Zh. Eksp. Teor. Fiz. 32, 920, 1957, Sov. Phys. JET Ρ 5, 749, 1957. [13] Igo, G. J., Clark, D. D., and Eisberg R. M., Statistical fluctuations in ionization by 31.5 MeV protons, Phys. Rev. 89, 879,1953. [14] Schultz, G., Charpak, G., and Sauli, F., Mobilities of positive ions in some gas mixtures used in proportional and drift chambers, Rev. Phys. appliquee, 12, 67, 1977. [15] Townsend, J., Electrons in gases, Hutchinson, London, 1947. [16] Price, W. J., Nuclear radiation detection Mc Graw-Hill, New York, 1958. [17] Rossi, B. and Staub, Η., Ionization chambers and counters, McGraw-Hill, New York, 1949. [18] Wilkinson, D. H., Ionization chambers and counters, Cambridge University press, Cambridge, 1950. [19] Fulbright, W., Ionization chambers in nuclear physics, Handbuch der Physik, SpringerVerlag, Berlin, 1958, vol 45. [20] Buneman, Ο., Cranshaw, Τ. Ε., and Harvey, J. Α., Design of grid ionization chamber, Can. J. Res., 27A, 191, 1949. [21] Naulin, F., Roy-Stephan Μ., and Kashi, Ε., Improved energy resolution in an ionization chamber through suppression of the electrical field distorsions, Nucl. Instr. and Meth. 180, 647, 1980.

Gas Filled Detectors

53

[22] Simon, G., Trochon, J., Brisard, F., and Signarbieux, C., Pulse-height defect in an ionization chamber investigated by cold fission measurements, Nucl. Instr. and Meth. A286, 220 1990. [23] Erskine, J. R., Braid, Τ. H., and Stoltzfus, J. C., An ionization chamber type of focal plane detector for heavy ions, Nucl. Instr. and Meth. 135, 67, 1976. [24] Tassan-Got, L., Stephan, C., Bizard, G., Duchon, J., Laville, J. L., L'Haridon, M., and Louvel, M., Heavy ion identification system for a magnetic spectrometer, Nucl. Instr. and Meth. 200, 280, 1982. [25] Saghai, B. and Roussel, P., Compteur proportionnel ä localisation pour la detection des ions lourds, Nucl. Instr. and Meth. 141, 93, 1977. [26] Rozet, J. P., Chetioui, Α., Piquemal, P., Vernhet, D., Wohrer, K., Stephan, C., and Tassan-Got, L., Charge-state distributions of few electron ions deduced from atomic cross sections, J. Phys. B. 22, 33, 1989. [27] Pfützner, Μ., Geissei, Η., Brohm, Th., Magel, Α., Münzenberg, G., Nickel, F., Scheidenberger, C., Schmidt, Κ. H., Sümmerer, K., Vieira, D., and Voss, B., Energy deposition of relativistic heavy ions in an ionization chamber at the FRS, GSI scientific report 1991, Darmstadt, 1992. [28] Badhwar, G. D., Calculation of the Vavilov distribution allowing for electron escape from the absorber, Nucl. Instr. and Meth. 109, 119, 1973. [29] Schapira, D., Teh, K., Blankenship, J., Burks, B., Foutch, L., Kim, H. J., Korolija, M., McConnell, J. W., Messick, M., Novotny, R., Rentsch, D., Shea, J., and Wieleczko, J. P., The HILI a multidetector system for light ions and heavy ions, Nucl. Instr. and Meth. A301, 76, 1991. [30] Sann, Η., Damjantschitsch, Η., Hebbard, D., Junge, J., Pelte, D., Povh, B., Schwalm, D., and Tran, Thoai, D. B., A position sensitive ionization chamber, Nucl. Instr. and Meth. 124, 509,1975. [31] Bocquet, J. P., Brissot, R., and Faust, H. R., A large ionization chamber for fission fragment nuclear charge identification at the Lohengrin spectrometer, Nucl. Instr. and Meth. A267, 466, 1988. [32] Gardes, D., Monnet, F., Barbey, S., Borderie, B., Dumail, M., Gobbi, Α., Rivet, M. F., and Volkov, P., a sectorized ionization chamber: the ΜΕΩ detection system, Nucl. Instr. and Meth. A247, 347,1986. [33] TP-MUSIC III- A new tracking detector for the ALADIN facility, Meijer, R. J., Liu Ζ. Α., Lühning, J., Lynen, U., Sann, Η., and Quick, W., GSI scientific report 1989, Dramstadt, 1990. [34] Gruhn, G. R., Binimi, M., Legrain, R„ Loveman, R., Pang, W., Roach, M., Scott, D. K., Shotter, Α., Symons, J., Wouters, J., Zisman, M., Devries, R., Peng, Y. C., and Sonheim, W., Bragg Curve spectroscopy, Nucl. Instr. and Meth. 196, 33, 1982. [35] Schiessl, C., Wagner, W„ Härtel, Κ., Kienle, P., Kvrner, H. J., Mayer, W„ and Rehm, Κ. E., A Bragg-curve spectroscopy detector, Nucl. Instr. and Meth. 192, 291,1982. [36] Asselineau, J. M., Duchon, J., L'Haridon, M., Mosrin, P., Regimbart, R., and Tamain, B., Performance of a Bragg curve detector for heavy ion identification, Nucl. Instr. and Meth. 204, 109, 1982. [37] Cebra, D. Α., Howden, S., Kam, J., Kataria, D., Maier, M., Nadasen, Α., Ogilvie, C. Α., Stone, N., Swan, D., Vander Molen, Α., Wilson, W. K., Winfield, J. S., Yurkon, J., and Westfall, G. D., Bragg curve spectroscopy in a 4π geometry, Nucl. Instr. and Meth. A300, 518, 1991. [38] Gramegna, F., Prete, G., Viesti, G., Iori, I., Moroni, A. and Ghinelli, F., Bragg curve spectroscopy at high rates, Nucl. Instr. and Meth. 243, 601, 1986. [39] Mittig, W. L., Gillibert, Α., Juzans, P., Schutz, Y., Stephan, C., Tassan-Got., L., and Villari, A. C. C., Unpublished data, 1989. [40] Shalev S. and Hopstone P., Empirical expressions for gas multiplication in 3 He Proportional counters, Nucl. Instr. and Meth. 155, 237,1978. [41] Diethorn, W., A methane proportional counter system for natural radiocarbon measurements, USAEC/NYO-6628,1956.

54

Claude Stephan

[42] Knoll, G. F., Radiation detection and measurement, Wiley and Sons, New York, USA, 1979. [43] Wolff, R. S., Gas constants for various gas mixtures, Nucl.Instr. and Meth. 115,461,1974. [44] Hendricks, R. W., Space charge effects in proportional counters, Rev. Sei. Instr. 40,1216, 1969. [45] Genz, H., Single electron detection in proportional gas counters, Nucl. Instr. and Meth. 112, 83,1973. [46] Va'vra, J., Wire chamber gases, Nucl. Instr. and Meth. A323, 34, 1992, and references therein. [47] English, W. N. and Hanna, G. C., Grid ionization chamber measurement of electron drift velocity in gas mixtures, Can J. Phys., 31, 768,1953. [48] Grunberg, C., Cohen, L., and Mathieu, L., Multiwire proportional and semiproportional counter with a variable sensitive volume, Nucl. Instr. and Meth. 78,102, 1970. [49] Hofmann, Th., Lynen, U., and Sann, Η., Gas purification for the TP-MUSIC detector, GSI scientific report 1991, Darmstadt, 1992. [50] Ford, Jr J. L. C., Position sensitive counters as focal plane detectors, Nucl. Instr. and Meth. 162, 277, 1979. [51] Borkovski, C. J. and Kopp, Μ. Κ., New type of position sensitive detectors of ionizing radiation using rise time measurement, Rev. Sei. Instr. 39, 1515, 1968. [52] Markham, R. G. and Robertson, R. G. H., High resolution position sensitive proportional counter, Nucl. Instr. and Meth. 129, 131, 1975. [53] Miller, G. L., Williams, N., Senator, Α., Stengaard, R., and Fischer, J., A position sensitive detector for a magnetic spectrograph, Nucl. Instr. and Meth. 91, 389,1971. [54] Fulbright, H. W., Markham, R. G., and Lanford, W. Α., position sensitive particle detectors used in a magnetic spectrometer, Nucl. Instr. and Meth. 108, 125, 1973. [55] Hafner, Η., and Duhm, Η. Η., The 1.4 m ΔΕ-Ε range position sensitive detector of the Heidelberg Q3D magnetic spectrograph, Nucl. Instr. and Meth. 160, 273, 1979. [56] Bianchi, L., Fernandez, B., Gastebois, J., Gillibert, Α., Mittig W., and Barrete J., SPEG: An energy loss spectrometer for GANIL, Nucl. Instr. and Meth. A276, 509, 1989. [57] Charpak, G., Boucher, R., Bressani, T., Favier, J., and Zupancic, C., The use of multiwire proportional counters to select and locate charged particles, Nucl. Instr. and Meth. 62, 262, 1968. [58] Boucher, R., Charpak, G. Dimcovski, Z., Fischer, G., Sauli, F., Coignet, G and Flügge, G., Investigation of some properties of multiwire proportional counters, Nucl. Instr. and Meth. 88, 149, 1970. [59] Erskine, G. Α., Electrostatic problems in multiwire proportional chambers, Nucl. Instr. and Meth. 105, 565, 1972. [60] Frieze, W., Dhawan, S., Disco, Α. Α., Fajardo, L., Majka, R., Marx, J. N., Nemethy, P., Sandweiss, J., and Slaughter, Α., A high resolution multiwire proportional chamber system, Nucl. Instr. and Meth. 136, 93, 1976. [61] Sauli, F., Limiting accuracies in multiwire proportional and drift chambers, Nucl. Instr. and Meth. 156,147, 1978. [62] Sanada, J., Growth of the avalanche around the anode wire in a gas counter, Nucl. Instr. and Meth. 196, 23, 1982. [63] Ford, T. D., Needham, G. Α., Brady, F. P., Romero, J. L., and Castadena, C. M., A multiwire chamber system for measurements of charged particle spectra, Nucl. Instr. and Meth. 228, 81, 1984. [64] Ball, G. C., Multiwire proportional chamber focal plane detector systems with digital readout, Nucl. Instr. and Meth. 162, 263, 1979. [65] Chalupka, Α., Bartl, W., Schönauer L., Bahnsen, Κ. U., Labedzki, H. J., Scheerer, H. J., Vonach, H., and Ziegler, G., New MWPCs for the Munich Q3D spectrograph, Nucl. Instr. and Meth. 217, 113, 1983. [66] Cunningham, R. Α., Sanderson, Ν. E., Snodgrass, W. N. J., Banes, D. W., Hoath, S. D., and Mo, J. N., Construction and performance of a multiparameter focal plane detector for use on a Q M G / 2 magnetic spectrometer, Nucl. Instr. and Meth. A234, 67, 1985.

Gas Filled Detectors

55

[67] Glässer, P., Rosier, H., Männer, R., and Specht, Η. J., A position sensitive multiwire proportional chamber for fission fragments with large solid angle, Nucl. Instr. and Meth. 141, 111, 1977. [68] Dangendorf, V., Bethge, K„ Kelbch, C , Kelbch, S , Ullrich, J., and Schmidt-Böcking, H„ Position sensitive gas proportional counter with good time resolution for low energy X-ray detection, Nucl. Instr. and Meth. A243, 465, 1986. [69] Dmitriev, G. D. and Frumkin, I. B., A study of M W P C accuracy of X-ray coordinate sensing at high counting rate, Nucl. Instr. and Meth. A295, 384, 1990. [70] Giorginis, G., Wochele, J., Kiontke, S., Maschuw, R., and Zeitnitz, Β., A three dimensional He-recoil M W P C for fast polarized neutrons, Nucl. Instr. and Meth. A251, 89, 1986. [71] Alekseev, G. D., Kalinina, Ν. Α., Karpukhin, V. V., Khazins, D. M., and Kruglov, V. V., Investigation of self quenching streamer discharge in a wire chamber, Nucl. Instr. and Meth. 177, 385, 1980. [72] Guillot, J., Morlet, M., Willis, Α., Denoit, M., Sellem, R., Beaunier, J., Bisson, Y., Chesneau, G., Goby, G., Lelong, P., Margaria, R., Maroni, Α., Menny, P., Seguy, D., Volkov, P., The SPES1 chambers at SATURNE, Private communication [73] Fuchi, Y., Tanaka, Μ. H., Kubono, S., Kawashima, H., Takaku, K., and Ichihara, T., Development of detectors for the second focal plane of SMART, R I K E N accel. Prog. rep. 25,129, 1991. [74] Villari, A. C. C„ Mittig, W., Blumenfeld, Y., Gillibert, A , Gangnant, P., and Garreau, L„ A test of new position sensitive detectors for SPEG, Nucl. Instr. and Meth. A281,240,1989. [75] Walenta, Η. Α., Heintze, J., and Schürlein, Β., The multiwire drift chamber, a new type of proportional wire chamber, Nucl. Instr. and Meth. 92, 373, 1971. [76] Atencio, L. G., Berg, G. P. Α., Von, Brentano P., Brinkmöller, Β., Hlawatsch, G., Meissburger, J., Moore, C. F., Morris, C. L., Paul, D., Römer, J. G. M., Rogge, M., Von Rossen, P., Sagefka, T., Seestrom-Morris, S. J., and Zemlo, L., The new focal plane detector for the magnet spectrometer Big Karl, Nucl. Instr. and Meth. A242,95,1985. [77] Engelage, J., Baumgartner, Μ. Ε., Beleal, Ε., Berman, Β. L., Bieser, F., Brady, F. P., Bronson, M., Carroll, J. B., Crawford, H. J., Flores, I., Greiner, D. E., Greiner, L., Hashimoto, O., Igo, G., Kadota, S., Kirk, P. N., Lindstrom, P. J., McParland, C., Nagamiya, S., Olson, D. L., Porter, J., Romero, J. L., Ruiz, C. L., Symons, T. J. M., Tanahita, I., Wada, R„ Webb, M. L„ Yamada, J., and Yee, H., Design of the BEVALAC heavy ion spectrometer system and its performance in studying 1 2 C fragmentation, Nucl. Instr. and Meth. A277, 431, 1989. [78] Henderson, R. S., Hausser, Ο., Hicks, Κ., Gunther, C., Faszer, W., Sawafta, R., and Poppe, Ν., Large area horizontal drift chambers for a focal plane Polarimeter at the T R I U M F medium resolution spectrometer, Nucl. Instr. and Meth. A254, 61, 1987. [79] Bertozzi, W., Hynes, Μ. V., Sargent, C. P., Creswell, C„ Dunn, P. C., Hirsch, Α., Leitch, Μ., Norum, Β., Rad, F. Ν., and Sasanuma, Τ., Focal plane instrumentation: a very high resolution M W P C system for inclined tracks. Nucl. Instr. and Meth. 141, 459, 1977. [80] Sjoreen, T. P., Ford, J. L. C. Jr, Blankenship J. L., Auble R. L., Bertrand F. Ε., Gross Ε. Ε., Hensley D. C., and Schull, D., The vertical drift chamber as a high resolution focal plane detector for heavy ions, Nucl. Instr. and Meth. 224, 421, 1984. [81] Charpak, G., Sauli, F., and Duinker, W., High-accuracy drift chambers and their use in strong magnetic fields, Nucl. Instr. and Meth. 108, 413, 1973. [82] Schroeder, L. S., Streamer chambers—their use for nuclear science experiments, Nucl. Instr. and Meth. 162, 395, 1979. [83] Bibber, K. v. and Sandoval, Α., Streamer chambers for heavy ions, Treatise on heavy ion science, vol 7, Plenum Press, New York, 1985. [84] Alard, J. P., Arnold, J., Augerat, J., Babinet, R., Bastid, N., Brochard, F., Costilhes, J. P., Crouau, M., De, Marco, N., Drouet, M., Dupieux, P., Fanet, H., Fodor, Z., Fraysse, L., Girard, J., Gorodetzky, P., Gösset, J., Laspalles, C., Lemaire, M. C., L'Höte, D., Lucas, B., Montarou, G., Papineau, Α., Parizet, M. J., Poitou, J., Racca, C., Schimmerling, W., Tamain, J. C., Terrien, Y., Valiro, J., and Valette, The Diogene 4π detector at Saturne Nucl. Instr. and Meth. 261, 379,1987.

56

Claude Stephan

[85] Mes, H., Anderson, H. L., Azuelos, G., Blecher, M., Bryman, D. A.Burnham, R. Α., Carter, A. L., Depommier, P., Dixit, Μ. S., Gotow, Κ., Hargrove, C. Κ., Hasinoff, Μ., Kessler, D., Leitch, M., MacDonald, J. Α., Mckee, R. J., Martin, J. P., Navon, I., Numao, T., Poutissou, J. M., Poutissou, R., Schlatter, P., Spuller, J., and Wright, C. S., Update on the T P C at T R I U M F , Nucl. Instr. and Meth. 225, 547, 1984. [86] Rai, G., Arthur, A , Bieser, F., Hamden, C. W., Jones, R„ Kleinfelder, S., Lee, K , Matis, H. S., Nakamura, M., McParland, C., Nesbitt, D., Odyniec, G., Olson, D., Pugh, H. G., Ritter, H. G., Symons, T. J. M., Wieman, H., Wright, Μ., Wright, R., and Rudge, R., A T P C detector for the study of high multiplicity heavy ion collisions, IEEE trans, on nucl. sei. 37, 56, 1990. [87] Beyerle, G., Glatz, Α., Gnirs, M., Herrmann, N., Jackel, W., Linke, R., Pelte, D., Schlesier, R., Trzaska, M., and Winkler, U., The central drift chamber of the 4π detector, GSI scientific report 1990, Darmstadt, 1991. [88] Daues, H. W., Fodor, Z., Gaiser, H., Gobbi, A , Grösch, H„ Reindl, M., Stelzer, H., Ero, J., Kecskemeti, J., Koncz, P., and Seres, Z., Status of the H E L I T R O N of the 4π detector, GSI scientific report 1990, Darmstadt, 1991. [89] Bagge, E. and Christiansen, J., The parallel plate counter as a self quenching particle measuring equipment, Naturwissenschaften 39, 298, 1952. [90] Breskin, Α., Progress in low-pressure gaseous detectors, Nucl. Instr. and Meth. 196, 11, 1982. [91] Sernicki, J., Timing properties of transmission avalanche counters at moderate specific ionization, Nucl. Instr. and Meth. A251, 81, 1986. [92] Stelzer, H., A large area parallel plate avalanche counter, Nucl. Instr. and Meth. 133,409, 1976. [93] Breskin, A. and Zwang, Ν., A fast bidimensional position sensitive parallel plate avalanche counter (PPAC) for light and heavy particles, IEEE Trans. Nucl. Sei., 25,126,1978. [94] Gaiardo, D., Lelong, P., Roussel, P., and Volkov, P., unpublished. [95] Gaiardo, D., Lelong, P., and Volkov, P., Delay lines in heavy ion detectors, IPNO-84-05, Orsay, 1984. [96] Hempel, G., Hopkins, F, and Schatz, G., Development of parallel plate avalanche counters for the detection of fission fragments, Nucl. Instr. and Meth. 131, 445, 1975. [97] Harrach, D. v. and Specht, Η. J., A square meter position sensitive parallel plate detector for heavy ions, Nucl. Instr. and Meth. 164,477,1979. [98] Kusterer, K., Betz, J., Harney, H. L., Heck, B., Liu Ken Pao, and Porto, F., A gas detector system for mass and charge identification of heavy ions, Nucl. Instr. and Meth. 177,485, 1980. [99] Butler, P. Α., Connel, Κ. Α., Burrows, J. D., El-Lawindy, A. M. Y„ Guidry, M. W„ James, A. N., Jones, G. D., Lauterbach, C., Morrison, T. P., and Simpson, J., The application of a multiple gas counter spectrometer to the study of heavy ion reactions, Nucl. Instr. and Meth. A239, 221, 1985. [100] Orr, Ν. Α., Mittig, W., Fifield, L. K., Lewitowicz, M , Plagnol, E., Schutz, Y , Zhan Wen Long, Bianchi, L., Gillibert, Α., Belezyorov, Α. V., Lukyanov, S. M., Penionzhkevich, Υ. E., Villari, A. C. C., Cunsolo, Α., Foti, Α., Audi, G., Stephan, C., and Tassan-Got, L., New mass measurements of neutron rich nuclei near Ν = 20, Phys. Lett. B258, 29, 1991. [101] Breskin, Α., A subnanosecond low pressure M W P C for heavily ionizing particles, Nucl. Instr. and Meth. 141, 505, 1977. [102] Binon, F., Bobyr, V. V., Duteil, P., Gouanere, M , Hugon, L., Spighel, M., and Stroot, J. P., Low pressure multiwire proportional chambers with high time resolution for strongly ionizing particles, Nucl. Instr. and Meth. 94, 27,1971. [103] Breskin, Α., Chechik, R., and Zwang, Ν., Heavy ion timing with very low pressure MWPCs, Nucl. Instr. and Meth. 165, 125,1979. [104] Golovatyuk, V. M., Ivanov, A. B., Nikitin, V. Α., Peshekhonov, V. D., and Zanevsky, Yu. V., A low pressure proportional chamber with a high space resolution, Nucl. Instr. and Meth. 145, 437, 1977.

Gas Filled Detectors

57

[105] Sauli, F., Applications of gaseous detectors in astrophysics, medicine and biology, Nucl. Instr. and Meth. A323,1, 1992. [106] Stelzer, Η., Μ ultiwire chambers with a two-stage gas amplification, Nucl. Instr. and Meth. A310, 103,1991. [107] Izycki, M., Martin, M., Rosselet, L., and Solomey, N., A large multistep avalanche chamber: description and performance, Nucl. Instr. and Meth. A310, 98, 1991. [108] Breskin, Α., Chechik, R., Fraenkel, Z., Jacobs, P., Tserruya, I., and Zwang, Ν., Low pressure multistep detector for very low energy heavy ions, Nucl. Instr. and Meth. 221, 363, 1984. [109] Angelini, F., Bellazini, R., Brez, Α., Massai, Μ. M., Spandre, G., Torquati, M. R., Bouclier, R., Gaudens, J., and Sauli, F., The microstrip gas chamber, Nucl. Phys. 23A, 254, 1991. [110] Oed, Α., Position sensitive detector with microstrip anode for electron multiplication with gases, Nucl. Instr. Meth. A263, 351, 1988. [111] Bouclier,R.,Florent, J. J.,Gaudaen, J., Millon,G.,Pasta, Α., Ropelewski,L.,Sauli, F.,and Shekhtman, L. I., High flux operation of microstrip gas chambers on glass and plastic supports, Nucl. Instr. Meth. A323, 240, 1992. [112] Barasch, E. F., Bowcock, T. J. V., Demroff, H. P., Elliot, S. M , Howe, M. R., Lee, B., Mazumdar, Τ. K., Mclntyre, P. M„ Pang, Y„ Smith, D. D„ Wahl, J., Wu, Y., Yue, W. K„ Gaedke, R. M., and Vanstraelen, G., Gas microstrip chambers, Nucl. Instr. Meth. A315, 170, 1992. [113] Angelini, F., Bellazini, R., Brez, Α., Decarolis, G., Magazzu, C., Massai, Μ. M., Spandre, G., and Torquati, M. R., Results from the first use of microstrip gas chambers in a high energy physics experiment, Nucl. Instr. Meth. A315, 21, 1992. [114] Angelini, F., Bellazini, R., Bosisio, L., Brez, Α., Massai, Μ. M., Spandre, G., and Torquati, M. R., A microstrip gas chamber on a silicon substrate, Nucl. Instr. Meth. A314,450,1992. [115] Biagi, S. F., Jackson, J. N., Jones, T. J., and Taylor, S., Initial investigations of the performance of a microstrip gas avalanche chamber fabricated on a thin silicon dioxide substrate, Nucl. Instr. Meth. A323, 258, 1992. [116] Nagae, T., Tanimori, T., Kobayashi, T., and Miyagi, T., Development of microstrip gas chambers with multi-chip technology, Nucl. Instr. Meth. A323, 236,1992. [117] Oed, Α., Geltenbort, P., and Budtz-Jorgensen, C., Substratum and layout parameters for microstrip anodes in gas detectors, Nucl. Instr. Meth. A310,95, 1991. [118] Budtz-Jorgensen, C., Bahnsen, Α., Olesen, C., Madsen, Μ. M., Jonasson, P., Schnopper, H. W., and Oed, Α., microstrip proportional counters for X-ray astronomy, Nucl. Instr. Meth. A310, 82, 1991. [119] Hartjes, F., Hendriksen, B., Schmitz, J., Schijlenburg, H., and Udo, F., Operation of the microstrip gas detector, Nucl. Instr. Meth. A310, 88, 1991. [120] Angelini, F., Bellazini, R., Brez, Α., Massai, Μ. M., Spandre, G., and Torquati, M. R., A microstrip gas avalanche chamber with two-dimensional readout, Nucl. Instr. Meth. A283, 755,1989. [121] Barbosa, A. F., Riekel, C., and Wattecamps, P., Two-dimensional X-ray detector based on microstrip and multiwire design, Nucl. Instr. Meth. A323, 247,1992. [122] Angelini, F., Bellazini, R., Bosisio, L., Brez, Α., Massai, Μ. M., Perret, Α., Spandre, G., and Torquati, M. R., A microstrip gas chamber with true two-dimensional and pixel readout, Nucl. Instr. Meth. A32, 229,1992. [123] Richter, J., Large size microstrip particle detectors, Nucl. Instr. Meth. A323, 263, 1992. [124] Proceeding of the Sixth International wire chamber conference, Vienna, 1992. Nucl. Instr. Meth. A323,1992.

2

Scintillation Detectors

Klaus D.

Hildenbrand

Table of Contents I. Introduction 59 II. Inorganic scintillators 61 A. Principles and general properties 61 B. Selected examples 62 III. Organic scintillators 68 A. General properties 68 B. Special classes of organic scintillators 69 IV. Glass scintillators 71 V. Scintillation in gases 71 VI. Light detection 72 A. Photomultiplier tubes, photodiodes 72 B. Sensitivity matching 73 C. Coupling principles 74 VII. Energy and time measurements in scintillators A. General aspects 74 B. Energy determination 76 C. Quenching 76 D. Time measurements 78 E. Mean timing 80 F. Particle identification 81 VIII. Concluding remarks 85 References 85

74

I. Introduction Any ionizing particle passing through matter undergoes energy loss in excitation and ionization processes. The same is true of γ-rays and neutrons, the latter interacting in an indirect way via (η,γ)- or (n,p)-reactions. With a certain probability the subsequent deexcitation processes in the medium lead to the emission of light. This phenomenon is called scintillation, and the scintillation light can be used as a means to detect the primary particle or photon and to measure its energy deposit in the medium. Scintillation is a special case of fluorescence which denotes the stimulation of light emission by any means whatsoever, e.g. by light of shorter wavelength. The very general nature of scintillation light production makes almost any material in any agregate state a scintillator. The question whether it is of practical use as a radiation detector depends, however, on a number of additional properties:

60

Klaus D. Hildenbrand

• The material must have a high scintillation efficiency, i.e. a high light output for a given energy deposited by the particle or photon. On an absolute scale the efficiencies are low; only a few percent of the deposited energy appear in the form of photons. • If to be used as an energy detector, the scintillator response i.e. the relation between light output and energy deposit should be as linear as possible; this requirement is only poorly met by many types. If the timing properties of the detector are more important, a material delivering light pulses with fast rise and decay time has to be chosen. In this case one might have to find compromises with the requirement of high scintillation efficiency or linearity. • The material has to be transparent for its own scintillation light. If this is not the case it can be only used in thin layers or films (ZnS being an example, used for light screens in electron-ray tubes). Transparency can also be achieved by combining the scintillating material with a socalled wave-length shifter; by absorption and reemission the primary light is shifted into a region of longer wave length, which is no longer self-absorbed. This trick can also be used to adopt the wave length better to the sensitivity of the electronic device used for the light detection such as a phototube or a light diode. • For practical application it has to be considered how well the material can be worked, in which sizes or forms it is available (important for inorganic crystals), and how easy or how difficult it is to handle (fragility, sensitivity against moisture or vapors, vacuum-resistivity etc.). The use of scintillation detectors is one of the oldest techniques in nuclear physics. In Rutherford's famous scattering experiments, α-particles, which were scattered on a thin Au-foil, were detected by counting by eye their light flashes on a scintillating screen. After the discovery of Nal(Tl) this material became the most popular scintillator in atomic and nuclear physics, especially used for the detection of γ-rays. Organic scintillators were mainly used for ß-counting and, in large arrays, in high energy physics. In the seventies various types of inorganic scintillators were newly developed or investigated in greater detail such as BGO, C W O , Csl, BaF 2 etc., which in the mean time have replaced N a l to a large extent. This development was stimulated mainly by medical applications such as of X-ray tomography and positron-emission tomography (PET) which depend on robust, cheap but highly dense scintillators containing heavy elements for effective photon detection. These investigations also led to a better understanding of the emission spectra of certain scintillators, and it was discovered that different light components of a material and their intensity ratio can be used for intrinsic particle (or particle-photon) discrimination. Atomic and nuclear physics did benefit from these developments. Beyond the traditional use as photon detectors, many of these scintillators are used as energy detectors for charged particles nowadays. Organic plastic scintillators, earlier used almost exclusively in high-energy physics, have also found wider applications. The trend towards higher accelerator energies especially in heavy ion nuclear physics requires thicker energy detectors, so semiconductor devices are prohibitively expensive in most cases. In addition, at higher energies, the traditional A£-£-identification method for charged particles is more and more

Scintillation D e t e c t o r s

61

replaced by A£-time-of-flight measurements, and scintillators are unbeatable in terms of time resolution. In this article we first describe characteristics of scintillator materials which are of practical importance in present-day nuclear and high-energy physics, starting with inorganic materials, then organic scintillators and finally, very briefly, glass and noble gas scintillators. Details on the scintillation process itself will be given only to the extent necessary to understand features of individual materials or of different classes of scintillators. Such details are found in the relevant literature [1,2] or in books on radiation detectors [3]. Material properties can be found in the catalogues of commercial companies we will refer to later.

II. Inorganic Scintillators A. Priniciples and general properties Scintillation processes in inorganic materials are closely related to their crystal structure [1,2], In insulators or semi-conductors with a band diagram as sketched in Fig. 1, the primary radiation ionizes the constituent atoms by promoting some of the electrons into the valence band. Electron-hole pairs are formed, the decay of which leads to the emission of photons. This situation - which is more complex in reality, since lattice interactions are of importance - is found only in pure, or intrinsic, crystals (left side in Fig. 1). Extrinsic materials, in which the pure host is doped with a small amount of a metallic impurity (right side in Fig. 1) have gained a much higher practical importance. G r o u n d state and excited state of these "activators" are located within the forbidden energy gap of the host crystal. The deexcitation of primary electron-hole pairs proceeds via the upper activator states into the activator ground states. The half-lives of these transitions are usually much longer than the migration times of the free electrons in the crystal; it is therefore the decay time of the activator, which determines the decay characteristics of the scintillation light. Since the final transition

Conduct Ion Band

Conduction Band

Act I v a t o r Excited Levels Photon Activator RadlatIon κ

χ

χ

κ

χ

χ

Valence Band

χ

Ground State

.

Valence Band

Fig. 1. Scintillation processes in inorganic materials are c o n n e c t e d w i t h the b a n d structure in the crystals. Left: Intrinsic crystal, right: Extrinsic (or d o p e d ) crystal

62

Klaus D. Hildenbrand

energy is smaller t h a n the g a p energy, the emitted light is shifted to longer wave lengths (compared to the p u r e crystal's scintillation light). This is often m o r e convenient for detection, and, m o r e i m p o r t a n t , the crystal becomes t r a n s p a r e n t for this light, since self-absorption is greatly reduced. T h e fact that the crystal properties are decisive for practically all features of inorganic scintillators leads directly to the following conclusions: • The concentration a n d distribution of the activator a t o m s are very i m p o r t a n t for the scintillator properties. • Small differences a m o n g various crystals may result in rather different physical properties such as light o u t p u t , decay times, energy resolution etc.. Such differences can be caused by crystal defects created during the p r o d u c t i o n and by impurities which act as additional activation centers or lead to m o r e light absorption. Tiny variations of the activator concentration over the volume can also cause a problem. • Any later change in the crystal structure will alter the characteristics. Such changes can be caused by mechanical or chemical impact, b u t also by radiation damage, which is especially i m p o r t a n t for applications in nuclear and high energy physics. • The scintillation properties will vary as a function of temperature. Higher crystal temperatures increase the probability of radiation-free deexcitation processes, in which the energy is dissipated as heat, hence causing a lower scintillation efficiency.

B. Selected examples In this section specific properties of c o m m o n l y used inorganic c o m p o u n d s are described. Table 1 summarizes their specific values.

1.

Nal(Tl)

A m o n g all inorganic crystals T1-doped N a l has the highest light o u t p u t (cf. Table. 1) and combines a rather linear energy response with a m o d e s t t e m p e r a t u r e dependence.

Table 1. Characteristic constants of common inorganic scintillators Parameter

Nal(Tl)

Density (g/cm2) 3.67 Emission Maxima (nm) 415 Relative Light Yield 100 Decay Constant (ns) 230 Refraction Index 1.85 Radiation Length (cm) 2.59

CsI(Tl)

BaF 2

4.51 4.88 550(330) 325(220) 45 20(2) 1000(0.5) 630(0.6) 1.80 1.50 1.86 2.03

BuGe^ 7.13 480 15-20 300 2.15 1.12

2

CaF 2

ZnS(Ag)

3.18 435 50 940 1.44

4.09 450 > 100 70 2.36

Values are taken from "The Scintillation Detectors Catalogue", the H A R S H A W Chemical Company. Emission Maxima and decay constants in parantheses refer to secondary intensity maxima, the refraction index is given for the main maximum wavelength.

Scintillation Detectors

63

The high light output and its historical importance have made it a standard material: light outputs of other inorganic scintillators are normally given in percentage o f t h a t of Nal(Tl). A certain disadvantage (which it shares with most of the inorganic materials) is the long decay time of 230 ns, which is unfavorable to high count rates; the moderate time resolution of a few ns, which can be achieved between two detectors, is directly connected with these slow pulses. In addition it shows a weak, but often unpleasant "after glow", a phosphorescence on a time scale of some ms. It can be grown in crystals of many liters volume, but the material is fragile and hygroscopic, so in air its properties will deteriorate very soon. Thus the crystals have to be encapsuled hermetically in air-tight housings. Since its discovery more than forty years ago [4] Nal(Tl) for decades has remained the most prominent material for photon detection, whereas it was not much used for particle detection. For nuclear spectroscopy, large γ-detector arrays have been built out of Nal(Tl), one example being the Darmstadt/Heidelberg Crystal Ball [5]. It comprises 162 modules which in in a close-to-4n-geometry surround an inner sphere of 50 cm diameter in form of a detector shell 20 cm thick. In high energy physics large devices were constructed as electrometric calorimeters, such as the Crystal Ball [6] at SLAC (later used at DESY), which consists of 672 modules of about 40 cm length also arranged in a spherical geometry. Such detector systems contain crystal volumes of many hundred liters. In recent years Nal(Tl) has been increasingly replaced by materials of higher density, which are described in the following chapter.

2.

CsI(Tl)

The light output of CsI(Tl) is about a factor of two lower than the one of Nal(Tl), with much longer decay times (cf. Table 1). Due to its higher density it has a larger γ-absorption coefficient and a shorter radiation length. In these respects it is surpassed only by 8 1 4 6 6 3 0 ! 2 , which has a much lower light output (or by C d W 0 4 and Z n W 0 4 which find application in medicine, but are not discussed here). The biggest advantage in comparison with N a l is that Csl is much less hygroscopic, so it can be kept in normal air for some time without any damage; over longer periods it should be stored in a dry atmosphere. These features made CsI(Tl) very popular as a detector material for high-energetic photons. In middle and high-energy physics it is used in shower detectors a n d / o r electromagnetic calorimeters. As an example, the C L E O II experiment at Cornell comprises 8000 crystals in a barrel-like geometry [7], The wide application it has found in nuclear physics in recent times is based on the fact that its different wave length components provide a very good intrinsic particle identification with a separation between γ-rays and charged particles and a m o n g those even between p,d,t and He isotopes. CsI(Tl) emits a small UV component at 330 nm with a shorter decay time of about 0.5 μβ. The main intensity is found in a broad distribution in the visible range, peaking at 550 nm, which is connected to the emission from the T1 activation centers. The UV component is due to the pure intrinsic deexcitation. A smaller T1 content decreases the

64

Klaus D. Hildenbrand

Δ»,

At2

Fig. 2. Intrinsic particle identification in CsI(Tl): As explained in the sketch on top, the scintillator puls is integrated in two different time windows (over 400 ns right after its beginning and over 1 μβ, starting 1.6 μβ later). When the corresponding integrals Ll and L2 are plotted against each other, the matrix shows separate lines for photons and the isotopes of H, He and Li. The cylindric crystal had 5 cm diameter, 3 cm length and was read out by a photomultiplier (from ref. 9)

visible c o m p o n e n t (and the total intensity), hence ultra-pure C s l emits preferentially in the U V region. It is a rather fast scintillator, since the long delay time (Table 1) is caused by the long-wavelength component. F o r a given T1 content the relative intensity of the UV c o m p o n e n t increases with the ionization of the exciting radiation [8]. The m e t h o d of intrinsic particle identification is explained [9] in Fig. 2: As indicated by the sketch, the typical CsI(Tl) light pulse is digitized within two separate windows. A short gate integrates the fast component, a wide one part of the slow component. When the corresponding quantities are plotted against each other as shown in the figure, the γ-rays, Η a n d H e isotopes show up as separated branches; heavier elements are barely resolved. T h e degree of separation depends on the details of the gate times chosen.

Scintillation Detectors

65

τ ϊ.π

2

Ρ

rl

0)

ZL

Lü

α

1 Ζ >

2

0 0

50

100

150

ErCsl cI (MeV) Fig. 3. Intrinsic particle identification in CsI(Tl) by means of the pulse rise time: When this time is plotted against the total detected light, as obtained by an integration over the full pulse length, a similar separation as the one in Fig. 2 is obtained. The abscissa is approximately calibrated in MeV. The detector had a volume of 3 χ 3 χ 5 c m 3 a n d w a s r e a d o u t b y a p h o t o d i o d e ( f r o m r e f . 10)

The m e t h o d requires a substantial electronic effort in multi-detector arrays. Therefore often an alternative method is applied, which makes use of the different pulse rise times of the products: Since the total light pulse is a superposition of the short and the long component, the pulse rise time depends on the relative yield of the two and allows for particle discrimination as well. T h e rise time is measured between the leading edge of the pulse and the zero-crossing point of a bipolar signal obtained by a suitable shaping of the original pulse. Figure 3 demonstrates the resolution obtained by this method [10], the particle separation being equivalent to the one seen in Fig. 2. The effect of the light o u t p u t in different wave length regions depending on the ionization density [ 1 , 8 ] is found in many inorganic crystals as well as in organic materials, and it is k n o w n since a long time t h a t it can be used for the discrimination of γ-rays from particles [11]. It is the additional possibility of separating the isotopes of light elements which has m a d e this pulse shape discrimination so p o p u l a r in recent years. CsI(Tl) is favoured because it provides the best separation. Nal(Tl) shows comparable results, but because of its inferior technical properties it is practically never used for this purpose.

3.

BaF2

O n e disadvantage of b o t h Nal(Tl) and CsI(Tl) is the slow rise and decay time, which is not very suited for fast timing. Time-of-flight measurement of the products, however, is

66

Klaus D. Hildenbrand

3 < Ν

iL t

m • V t

Ol Ο η

ζ

w i d e g a t e BaF^» ( a . u . ) Fig. 4. Particle identification in a 25 cm long B a F 2 crystal [14,15]. T h e short integral over the fast light component is plotted against the total light (cf. sketch in Fig. 2). The main branches show neutral products (photons and neutrons) and the isotopes of Η and He

the only method to discriminate γ-rays from neutrons effectively, because in many cases the described pulse shape discrimination is insufficient for this purpose. Hence many investigations of scintillators have searched for fast light components which could help to overcome the problem. This search was successful in the case of BaF 2 , in which a fast UV component around 220 nm with only 0.6 ns decay time was found [12]. By integrating the two light components separately (as described in Fig. 2 for CsI(Tl)) an intrinsic particle discrimination is feasible [13]. The material itself can be handled without any problems. It is insensitive to severe outer conditions and hard against radiation, can easily be manufactured and be grown to fairly large sizes. These aspects have made BaF 2 , despite its low light output, a favourite material for large-size γ-detector devices: The photon spectrometer T A P S [14] comprises 384 hexagonal crystals of 25 cm length. Figure 4 displays the particle identification achieved [15] in one of these crystals; the fast is plotted vs. the total light component. It is worse than the one in Fig. 2, but good enough to discriminate the neutral products (neutrons and photons) from the charged ones; the very good time resolution of at least 100 ps then allows neutrons to be discriminated from photons through their time of flight, as demonstrated [15] in Fig. 5.

Scintillation Detectors

5

10

15

20

25

30

67

35

ET (MeV) Fig. 5. The - for inorganic scintillators - very good timing properties of BaF 2 allow a neutronphoton discrimination by time of flight measurement; this is demonstrated in the figure which shows a plot TOF vs. deposited energy. Displayed are the neutral events of Fig. 4 after a window has been set on the photon/neutron branch (from ref. 15)

4.

Bi4Ge3Ol2

B i s m u t h g e r m a n a t e , b e t t e r k n o w n as B G O , h a s t h e h i g h e s t d e n s i t y a m o n g all i n o r g a n i c scintillators of p r a c t i c a l i m p o r t a n c e ; d u e t o its h i g h effective e l e m e n t n u m b e r it h a s a sizable p h o t o p e a k efficiency a n d γ - r a y a b s o r p t i o n . T h e s e f e a t u r e s , h o w e v e r , g o a l o n g with a very m o d e s t light o u t p u t of less t h a n 1 0 % o f t h a t of N a l ( T l ) . T h i s l e a d s t o a n o n l y m o d e r a t e e n e r g y r e s o l u t i o n , so t h e m a t e r i a l is m a i n l y used in c a s e t h e γ - r a y d e t e c t i o n efficiency is t h e decisive r e q u i r e m e n t . T h i s is w h y B G O finds its m a i n a p p l i c a t i o n in C o m p t o n - s u p p r e s s i o n shields f o r h i g h - r e s o l u t i o n G e r m a n i u m γ - d e t e c t o r s . In a n o p t i m i z e d g e o m e t r y t h i s shield, p u t t o g e t h e r f r o m single B G O pieces w i t h s e p a r a t e r e a d - o u t , s u r r o u n d s t h e G e c r y s t a l as c o m p l e t e l y as p o s s i b l e (Fig. 6). By a n t i c o i n c i d e n c e w i t h t h e G e - d e t e c t o r t h e C o m p t o n - s c a t t e r e d γ - r a y s a r e s u p p r e s s e d , t h u s p r e p a r i n g t h e p h o t o p e a k s in t h e h i g h r e s o l u t i o n s p e c t r u m m e a s u r e d with t h e G e - d e t e c t o r . D e t e c t o r a r r a y s with l a r g e n u m b e r s of s u c h B G O - s u p p r e s s e d G e - d e t e c t o r s a r o u n d t h e t a r g e t a r e in u s e in several l a b o r a t o r i e s , p r e s e n t i n g t h e s t a t e - o f - t h e - a r t t e c h n o l o g y in γ - s p e c t r o s c o p y . T h e m o s t a m b i t i o u s p r o j e c t s of this k i n d a r e G A M M A S P H E R E [ 1 6 ] a n d E U R O G A M [17], w h i c h p r o p o s e c o m b i n i n g u p t o 70 h i g h - r e s o l u t i o n d e t e c t o r s in a c l o s e - t o 4TC-geometry. In h i g h - e n e r g y physics, B G O is b e i n g u s e d in e l e c t r o m a g n e t i c c a l o r i m e t e r s . T h e m o s t i m p r e s s i v e e x a m p l e is t h e c e n t r a l B G O b a r r e l of t h e L 3 e x p e r i m e n t [ 8 ] at C E R N , w h i c h c o m p r i s e s a l m o s t 12000 single crystals.

68

Klaus D. Hildenbrand

PM-Tube

Fig. 6. Schematic cross section through a C o m p t o n suppressed Ge-detector. The G e crystal is surrounded almost completely by a shield of several B G O scintillators, which are read out by individual phototubes. By anti-coincidence the photons which are Compton-scattered out of the center crystal are suppressed in the spectrum of the Ge-detector

5. CaF2 ( Eu) C a F 2 , activated with Eu, is a very inert inorganic scintillator, in this respect comparable to BaF 2 . It has a fairly high light output and is sometimes used as a long-decay-time material in socalled phoswich combinations (see Sect. VII. F).

6.

ZnS(Ag)

Ag-activated ZnS is included here because of a peculiar feature: It cannot be grown in larger crystals but only as a polycrystalline powder, hence it can only be used in thin layers. In thicker layers it is not transparent for its own light. It is one of the oldest scintillators known with a very high light output; since the early days of scintillation detectors it has found application on scintillating screens and in cathode ray tubes. ZnS(Ag) powder imbedded in polymerized plastic material is a very exotic scintillator. It shows a light output comparable to ZnS, and is often used for neutroncounting; for this purpose sometimes a few percent of 6 Li are added.

III. Organic Scintillators A. General properties Scintillation in organic materials, in contrast to inorganic scintillators, is not an effect of the crystal lattice; it is based on the excitation of individual molecules, and in principle it does not matter whether the medium is a gas, a liquid or a solid. This implies the following features:

Scintillation Detectors

69

• The scintillating molecules can easily be imbedded in other materials, dissolved in liquids or mixed with a polymerizing substance. The choice of these secondary materials is mainly determined by practical i.e. technical considerations and of course limited to those which have a good transparency for the light to be transmitted. • Special features, such as the decay time or the scintillation efficiency will depend - if at all - only very weakly on the temperature. Despite this obvious simplicity the details of excitation and deexcitation are of complex nature and can be found elsewhere [ 1,3]. All organic c o m p o u n d s used contain symmetric structures (benzene rings) with pronounced singlet (S) and triplett (T) level schemes on which vibrational states are built. After the primary excitation the higher-lying S-states decay very fast, preferentially radiationless, to the first excited S!-state, and it is its decay back to the S 0 -ground state (or in its vibrational states) which delivers the main scintillation component. The lifetime of S j is typically a few ns; so - with a few exceptions - organic scintillators are much faster than inorganic crystals. Nevertheless also slower components are present in certain scintillators. These are ascribed to deexcitations of molecular triplett states, which are excited by conversion of the primarily excited S-states, but in most cases this phosphorescence is very weak. In analogy to the situation in inorganic crystals the ratio of the two components depends on the ionization density and might hence be used for internal particle discrimination and here especially for γ-neutron separation [11]. F o r further discrimination between different particles the effect is too weak and cannot compete with e.g. the internal discrimination in CsI(Tl). Although the deexcitation energy of the main component (decay Sj to S 0 ) is smaller than the average energy amount in excitation (which also populates levels above Sj) the materials are not too transparent. That is why all organic scintillators additionally contain a wave-length shifting agent which absorbs the primary UV light and emits it again in the visible light region (cf. Sect. VI). Since organic scintillators are very often of sizable dimensions, the so-called bulk attenuation length is an important parameter. It denotes the distance in an infinitely big piece of scintillator over which the intensity drops to 1/e of its primary value. The effective attenuation length in real devices can be much shorter, determined by the dimensions and the quality of the surfaces, since the light undergoes many reflections with unavoidable losses before reaching the photo-sensitive device.

B. Special classes of organic scintillators The large variety of scintillating organic compounds together with the choice of the wave-length shifter and of possible liquid or solid solvents has led to an immense selection of scintillators. The description of all of them would go far beyond the scope of this article. In the following we briefly introduce the most important classes of scintillators, as far as practical applications in nuclear and in high-energy physics are concerned. For the same reason Table 2 presents characteristic data of only a few scintillators [19,20] to demonstrate the bandwidth of possible properties.

70

Klaus D. Hildenbrand

Table 2. Selection of commercial organic scintillators Parameter

Anthracene

BC400 NE102A

BC418 Pilot U

BC444 NE115

BC428

Relative Light Yield Bulk Atten. Length (cm) Emission M a x i m u m (nm) Decay Constant (ns)

100

65 250 423 2.4

67 100 391 1.4

41 180 428 180

50 330 490 12

447 30

BC430

45 580 16.8

Values taken from producers catalogues [9,20]. Scintillators with " B C " denote t r a d e n a m e s of B I C R O N C o r p o r a t i o n , the other names are identical materials from N E Technology, Ltd. The density of the plastic scintillators (all entries except anthracene) is 1.03 g/cm 3 , their refraction index 1.58. In terms of light yield, Nal(Tl), the standard of the inorganic scintillators in Table 1, would have 230 in the present table.

1. Pure organic

crystals

Only a few organic compounds have found broader application in form of crystals, the best known being anthracene and stilbene. This is not in conflict with what has been said before about the nature of organic scintillation: Both compounds d o scintillate as vapours too, but in practice they have been used only as crystals, mainly for neutron and beta-counting. Nowadays they would only be of historical interest, if anthracene did not hold the record in scintillation efficiency among all organic scintillators. That is why the efficiency of organic scintillators is specified in relative units with anthracene as a basis. The light output of Nal(Tl), which is the standard for inorganic scintillators, is about a factor of 2.3 higher than that of anthracene, which allows the values of Table 1 and 2 to be compared on a common scale.

2. Plastic

scintillators

In terms of the quantities produced, plastic scintillators certainly are the most important class of organic scintillators used in nuclear or high-energy physics and in radiation monitoring. Together with the wave-length shifter the primary organic scintillating agent is dissolved in a solvent which can be polymerized, e.g. aromatic solvents such as vinyltoluene and styrene. A wide variety of such scintillators is available, tailored to different needs concerning decay times or attenuation lengths (see Table 2). The materials are rather inexpensive and produced in all different forms, available as cylinders and rods of sizable dimensions, as fibers of sub-millimeter diameter and foils/plates from 50 μιη to 5 cm thickness. They can be easily cut or tooled, but because of their softness (softening points about 75 °C) one has to apply certain techniques e.g. diamond machining, in order to obtain high-quality surfaces. After a gentle heating the material can also be bent or pressed to any thickness. Dissolved in certain organic solvents one can prepare thin films down to the few ten μg/cm 2 range [21].

Scintillation Detectors

71

There is another very similar class of scintillators, where aliphatic c o m p o u n d s (acrylic glass) are used as polymerizing solvents. This leads to harder materials which are easier to work and less sensitive to mechanical or chemical stress. Their light output, however, is only about half of that of the scintillators in aromatic solvents, so they are mainly used for applications, where severe outer conditions need such stable and firm compounds.

3. Loaded organic

scintillators

Organic materials are specially suited for the detection of charged particles; since they contain only light elements their response to γ-rays and neutrons (via n, γ-reactions) is very weak, and the interaction occurs by C o m p t o n scattering rather than by photoelectron emission. Therefore attempts have been made to add heavy elements such as P b or Sn to the plastic material, which leads, however, to a decrease in light output. In terms of achievable energy resolution such scintillators are clearly inferior to inorganic crystals, but they offer the advantages of their class, namely a fast time response and large volumina at moderate cost. A higher light output is found in liquid scintillators which are loaded with an isotope with large cross section for neutron capture with subsequent charged-particle - or γ-emission (e.g. 10 B(n, a) or Gd(n, γ)) which find application in neutron counting. Details about these and similar scintillators are to be found in the catalogues of commercial companies [9,20].

IV. Glass scintillators Scintillation is also found in a variety of glasses doped with certain activators such as cerium. The light output of these scintillating glasses is typically 20 to 3 0 % of anthracene only; the decay times are a some ten ns and hence intermediate between organic and inorganic scintillators. The material is very insensitive to mechanical and chemical stress, high tempertaure or high radiation flux, and finds application in places where such extreme conditions exclude other classes of scintillators. Special glasses have been developed for neutron counting which are loaded with a few percent of 6 Li.

V. Scintillation in gases Many gases show scintillation, but in practice only noble gases have found a wider application. The excited single atoms emit spectral lines and continua preferentially in the UV region; fast components of ns decay time allow for good time resolutions. Since the gases are filled in containers, active volumes of all sizes and shapes are easily built. Very attractive are liquid gases: Liquid Xe has a density of 3 g/cm 3 and hence a sizable stopping power. Its total light output is comparable to that of plastic scintillators and it

72

Klaus D. Hildenbrand

seems to be well suited for large-volume calorimeters in high-energy physics [22]. Xe exhibits an unusual linearity in light response, quenching ( c f . section VII.C.) seems to be much weaker than in any other scintillator [23],

VI. Light detection A. Photomultiplier tubes, photodiodes The achievable energy and time resolution in scintillation detectors depend to some extent on the properties of the devices which are used to record the very weak scintillation light. There is a choice between two technical possibilities: Photomultiplier-tubes (PMTs) or semiconductor photodiodes. In both devices the incident photons release electrons from a photosensitive layer. In the P M T this is the thin photocathode at the inner surface of the glass bulb, in a diode it is the silicon layer between the electrodes. In the case of the diode the corresponding charge is read out directly, in the P M T the electrons are accelerated through a system of electrodes, so-called dynodes, on the surface of which each electron releases a certain number of secondary electrons, until the final charge pulse is taken off the anode. There is a huge variety of both P M T s and photodiodes on the market; the best survey together with descriptions of basic working principles can be found in handbooks or catalogues of commercial companies [24], Here we will present only some considerations which are important in connection with scintillator read-out. P M T s offer high gain factors (up to 107) and fast pulse rise times (down to and below 1 ns). The so-called linear-focussed tubes show the best performance in this respect. They do, on the other hand, suffer significantly from magnetic fields and have to be shielded even against the earth's magnetic field, so their use in very strong fields is very complicated and practically not feasible. There are other types of a more simple geometrical design which can, at the expense of achievable rise time be operated in rather high fields, at least in a certain orientation with respect to the field axis. One disadvantage common to all P M T s is the exponential dependence of their gain factor on the applied high voltage, so they need highly stable supplies, but even then the gain may drift for various reasons. Depending on the requirements of the experiment these variations have to be controlled by stable reference-light sources, e.g. by radioactive sources with γ- or α-lines which are recorded in the scintillator itself or stable light flashes provided by light-emitting diodes or a laser [28]. Finally P M T s suffer from pulse-height and pulse-form variations (the latter being dangerous for timing) at high count rates and/or high-intensity pulses, caused in both cases by the high space charge (high current) in the dynode system. The effect can be reduced by a special design of the tube base which supplies the high voltage to the dynodes. Photodiodes have a gain factor of one, so they need a preamplifier for read-out, followed, if necessary, by a shaping main amplifier. As semiconductor devices they have the intrinsic problem of noise, which limits their size to a few cm 2 of active area, if the noise level is to stay below an acceptable limit. O n the other hand, they require much

Scintillation Detectors

73

less space than a P M T , can be run in magnetic fields and are cheaper, so it can pay to use more than one diode for the read-out of a large scintillator. Their rise time cannot compete with that of PMTs; that is why they are mainly used in set-ups where the timing properties are not decisive, especially in connection with inorganic scintillators, the slow rise time of which do not allow for fast timing.

B. Sensitivity matching O n e important condition for effective light conversion is the matching between the emission wavelength of the respective scintillator (cf. values given in Tables 1 and 2) and the sensitive range of the P M T cathode or the diode. Figure 7 shows that the sensitivity curve of a P M T , which does not vary too much between different types, matches well with the emission wavelength of Nal(Tl) which is very close to the one of a blue-emitting plastic scintillator. CsI(Tl) and the green or orange-emitting plastic scintillators (specially developed for this purpose) are much better covered by a photodiode. If light below 380 nm has to be detected (for example from BaF 2 ) so-called UV-extended tubes or diodes have to be used. In the case of P M T s this means entrance windows of quartz glass (which considerably increases the price) or the use of special transparent protective layers on the diodes. A possible method of overcoming bad matching is in the use of a wave length shifter, which is placed between the scintillator and the diode. It consists of a fluorescent

="

1 ,o

1 .0

0.5

0.5

200

400

600 WAVELENGTH

800

(nm)

Fig. 7. Emission curves of CsI(Tl) and of Nal(Tl); the latter is very similar to the one of a blueemitting plastic scintillator (cf. Tables 1 and 2). F o r comparison the sensitivity of a photomultiplier tube (bi-alkali photocathode) and of a photodiode are shown qualitatively by the solid curves; their extension by dashed curves refer to U V or IR extended devices (values taken from ref. 24)

74

Klaus D. Hildenbrand

material which absorbs light of a certain wave length, normally in the UV, and remits it at a longer wave-length. Such shifters are available in a large variety of absorption and emission ranges, mainly as acrylic glasses in which the fluorescent agent is incorporated.

C. Coupling principles Whereas all surfaces of a scintillator have to be prepared for light reflection as completely as possible (cf. Sect. VII. A) the surface window, through which the light is to be read out, has to have a high transparency. This requires a good matching of the refraction indices of all materials involved such as the scintillator, glass of the P M T or the cover epoxy of the diode, a possible light guide, but also the optical cement, glue or silicon grease with which these parts are fixed together. This is rather easy in the case of plastic scintillators, which all have η ~ 1.58, a value close to the index of most glasses or epoxy resins. For inorganic crystals with their widely varying indices (cf. Table 1) a good matching is less trivial. In practice it is often necessary to use a light guide between the scintillator and the read-out device. This always leads to a loss of light, which has to be tolerated in cases where e.g. a P M T cannot be mounted directly for spatial reasons or because it has to be operated outside a magnetic field in which the scintillator is placed. Light guides are also needed if the surfaces of scintillator and tube are different in shape. Light transport follows Liouville's theorem, which states the light flux per unit area and solid angle being a constant. Hence a light guide of diminishing cross section, which couples a large scintillator surface to a smaller P M T window, can at best transport an intensity reduced by the ratio of the surfaces. If the surfaces are equal but very different in shape, e.g. in case of a long rectangular scintillator connected to a round P M T , one can use a so-called twisted guide. It has the cross section of the scintillator, but is cut from the opposite side into several strips like a comb. Then the single strips are gently bent and led together in the geometry of the photocathode. In this configuration the light intensity is only reduced by the unavoidable losses during the multiple reflections, which depend strongly on the surface quality.

VII. Energy and time measurement in scintillators A. General aspects Before turning to the measurement of energy and time with scintillators we have to mention some general differences in their response to various species such as photons, neutrons or charged particles and some different mechanisms present at low or high incident energies. The qualitative response to charged particles is relatively simple in all materials, since they deposit their kinetic energy almost quantitatively by ionization. The

Scintillation Detectors

75

conversion into light, however, is strongly dependent on the ionization density and hence on the specific energy loss d£/dx, leading to a nonlinear relation between the produced light and the energy. This phenomenon called "quenching" is described in a separate chapter below. The energy of γ-rays can only be measured in scintillators with a sizeable photopeak efficiency, i.e. high-Z materials. Therefore organic materials (except the loaded scintillators as described in Sect. Ill) are usable for the detection but not for the energy determination of photons. This is the domain of inorganic crystals with their high photopeak efficiency. The mechanism of γ-ray detection changes as the photon energy becomes greater that 1 MeV. The single-photoelectron process vanishes in favour of the formation of an electromagnetic shower. It starts with e + /e~ -pair production by the incident photon, followed by an avalanche of annihilation of pair-production processes. The total light emitted is given by the contribution from all charged particles created. The effectivity of a material for forming such a shower and hence for high-energy photon detection is expressed by the radiation length x 0 (cf. Table 1); it describes the distance over which a high-energetic electron loses all but a 1/e fraction of its energy. For an effective recording of the total energy one needs fairly long (and expensive) crystals, typically ~ 1 5 x 0 for photons of several hundred MeV. The shower also develops a certain transversal width, typically of a few cm as expressed by the so-called Moliere radius. Instead on one huge crystal one therefore prefers to use an array of long crystals, packed side by side as closely together as possible, which are read-out separately through their back surfaces [14]. Their front face width is chosen somewhat larger than the double Moliere radius, i.e. the mean shower diameter. The total photon energy is obtained by adding the signals from the neighbouring crystals into which the shower has "leaked". By weighting these contributions the position of incidence onto the center crystal can be determined with a resolution better than the module's radial dimensions. In general the response of such an electromagnetic shower detector of given material and geometry can be reliably simulated [25] which is necessary for the energy calibration. Neutrons are detected in scintillators via nuclear reactions such as (η,γ), (η,ρ) or (η,α). At higher incident energies more complicated reactions play an additional role. That explains why the efficiency of neutron detection, i.e. the probability of seeing the neutron at all depends in a complicated way on the scintillator used and its composition as well as on the neutron energy. For acceptable efficiencies one needs fairly big detectors, and except for a few special applications, one uses organic scintillators for neutron detection. Even then the efficiency has to be determined by separate measurements or by Monte-Carlo simulations [26], The efficiency can be increased by slowing down the neutrons in suitable absorbers (or in the detector itself, which again needs big volumina, e.g. a liquid scintillator tank). At energies above a few hundred MeV it can be increased by using thin scintillators sandwiched between iron sheets [27], in which it is highly probable that each incoming neutron will create a hadronic shower of charged particles. The complicated response of neutrons makes a direct energy measurement hardly feasible, so in most cases the energy of neutrons is determined by a time-of-flight measurement.

76

Klaus D. Hildenbrand

B. Energy determination An effective energy determination requires as complete a light collection as possible. Since the light is emitted isotropically from all parts of the ionization track, the detector surfaces have to be reflective. Otherwise only a fraction of the total light would be detected, namely that part which reaches the read-out device directly or after internal reflection on the surrounding walls. The best results are not obtained with highly reflective specular surfaces like polished metals, but with diffuse materials into which the scintillator is packed. Such materials are, e.g. white metal oxides such as aluminium, magnesium or titanium oxide, either as powder or as suspension which can be painted on the surfaces ("reflective paint"); good results also are achieved with aluminium foil and teflon tape which is wrapped around the scintillator. In general the results are better the more opaque the reflecting surfaces are. The obtainable energy resolution is affected by the light collection efficiency in two distinct ways. Firstly, via the photon statistics, i.e. the number of collected photons; a loss of photons will worsen the resolution. The second feature is the degree of uniformity of light collection from different sections of the scintillator. It can be tested by separately irradiating certain regions with a collimated radioactive source or an accelerator beam. A perfect uniformity would result in an equal mean light signal irrespective of the position of incidence. In practice this is not easy to achieve and especially in large scintillators, the position dependence may be the limiting factor in the energy resolution, so the resolution depends significantly on the size and the shape of a detector. To some extent the dependence can be corrected by a special surface treatment, a controlled depolishing or different wrapping of certain surface regions. It is not possible to quote general quantitative figures for the obtainable energy resolutions in scintillators, since the bandwidths among particles and photons, organic and inorganic materials and various energy regimes are too large. Another factor limiting the energy resolution is the low total light efficiency of scintillators: In Nal(Tl), the material with the highest efficiency, only about 7 % of the energy deposited appears in the form of scintillation light, resulting in about 4.10 4 primary photons per 1 MeV of energy loss of a minimum ionizing particle. Values in other scintillators are lower according to their relative light yields as given in Tables 1 and 2. The decisive quantity in terms of statistics, the number of photoelectrons in the P M T , is at least a factor 10 lower; this loss is caused by the collection efficiency and photosensitivity of the P M T cathode which is typically about 20%. This leads to very small electron numbers, and the resolution is determined by Poisson statistics to σ(Ε)/Ε = 1/yfN. In view of these poor statistics, scintillators, e.g. in γ-spectroscopy, can never compete with semiconductor devices: For a γ-line of 1 MeV energy in CsI(Tl), resolutions of 4 to 8 % are obtained, compared to a few permille in a Ge dedetctor.

C. Quenching We have seen that the total light output is very low in all scintillators. There are some effects which may decrease it even further; they are globally summarized by the expression "quenching".

Scintillation Detectors

77

Fig. 8. Light output (arb. units) of the plastic scintillator N E 102A (cf. Table 2) for various light charged particles. Only electrons exhibit a linear relation between the deposited energy and the light output (from ref. 20)

The so-called concentration quenching is a particularly unpleasant feature. The light output is at a maximum for a certain concentration of emitting centres, the activators in inorganic crystals or the primary scintillating molecules in organic scintillators. At any given concentration the output strongly depends on the energy amount deposited over a certain distance, i.e. on dE/dx. Should this quantity become too high, the number of scintillating centres will be too small and an increasing fraction of the energy is not converted into light. This leads to a non-linear response: The curve of light output vs. deposited energy is not a straight line. The dependence on dE/dx explains why this quenching is stronger the slower and/or the heavier the particles are. As shown in Fig. 8, only electrons are not affected, they show a linear response. Protons deviate from this ideal behaviour [20], At an energy of 100 MeV the proton signal is 20% smaller; at very high energies the discrepancy is reduced, but the value always stays below the corresponding electron figure. For a quantitative comparison of different scintillators the ratio of the light-output of α-particles to that of electrons is often quoted. For heavier nuclei the situation is even worse [29]. Figure 9 displays the situation for elements ranging from Η to Br. Attempts have been made to describe these figures quantitatively (cf. ref. 29 and references therein); most approaches, however, try to parametrize measured values, and the predictive power is not too high. A rather realistic formulation [1] assumes the following relation between dL/dx and d £ / d x (specific light emission vs. specific energy loss): dL/dx oc

dE/dx/ll-KB{dL/dx)1

(1)

78

Klaus D. Hildenbrand

Fig. 9. Response of a fast plastic scintillator to elements ranging from Η to Br (from ref. 29). The curves demonstrate both the non-linear relation between deposited energy and light output for each element and the drastic absolute decrease in light intensity for the heavy elements; both effects are due to quenching (see text for details)

where the non-quenching limit is given by the parameter KB going to zero. The light output is given by L=

^•dx dx

(2)

integrated over the range of the particle. Satisfactory agreement is found for not too heavy masses with experimental XB-parameters as given in ref. 30.

D. Time measurement A unique feature of many organic scintillators is their fast time response, hence they are ideally suited for time-of-flight measurements. Analogous to the energy determination the photon statistics determine the resolution achievable, but here in addition the

Scintillation D e t e c t o r s

79

photon time distribution is of importance. This distribution is primarily given by the length of the scintillator light pulse (rise and decay time), so for high timing requirements one uses very fast scintillators and PMTs. However, in large counters the geometry is again the limiting factor: Light velocities are of the order of 20 cm/ns, so in a 5 cm long scintillator, which is read out at one end, the arrival times of the light pulses at the P M T will vary by 250 ps depending on the position of incidence. This problem can be solved by the mean timing method (c/. next section). The transit-time spread through the P M T and effects of the electronics used can be additional sources of degradation of the resolution, but these aspects shall not be treated here. If the irradiated area is small enough, time resolutions (σ) of the order of 30 ps are not too difficult to obtain with minimum ionizing particles. The variation as a function of the deposited energy in 2 cm thick, long scintillators and hence the light intensity is depicted in Fig. 10: It shows the time resolution achieved with 512 strips of a scintillator wall [28], designed for A£-TOF identification of particles (cf. Sect. F). Time is measured in these strips by mean timing as described in the next section. The resolution σ(Τ) achieved is 105 ps for minimum ionizing particles. With increasing light intensity it becomes better (by a factor roughly proportional to the square root of the intensity)

w c

1

10

10

Light Intensity [ M I L ] Fig. 10. T i m e resolution σ(Τ), o b t a i n e d with 512 scintillator strips of varying length, m e a s u r e d with a laser as f u n c t i o n of the p r i m a r y light intensity (from ref. 28). T h e latter is given in units of M I L , 1 M I L being t h e light p r o d u c e d by a m i n i m u m ionizing electron. P l o t t e d is the m e a n resolution of all 512 strips, d r o p p i n g with improving p h o t o n statistics f r o m 105 ps for 1 M I L to a b o u t 40 ps at a h u n d r e d times m o r e light. T h e e r r o r b a r s d e n o t e the variance f o u n d a m o n g the strips; it is mainly d e t e r m i n e d by the different lengths, which vary between 45 c m a n d 165 cm, i.e. by the c h a n g i n g p h o t o n statistics

80

Klaus D. Hildenbrand

HV

HV —TS

L

PUT

TL

I / / / / Z PMT ——

'////λ

L

—

X

—

J —

R

TR

s

Fig. 11. Sketch of a long scintillator strip which is read out at both ends by a PMT. From each tube an energy signal Ε and a time signal Τ is derived. By means of these quantities one can, by averaging methods, derive the primary light L 0 and the time of flight TOF independent of the position of incidence (for details see text)

and finally levels off at about 40 ps, demonstrating that at high intensities the photon statistics is no longer decisive.

E. Mean timing The position dependence of energy and time signals described can be overcome by a double read out of a scintillator. Consider a situation, as sketched in Fig. 11: A scintillator, the length of which is large compared to its lateral dimensions (height, width) is read out at both ends via a P M T . The signal from each P M T is split into two signals for energy (E) and time (T), so the four signals EL, EK, TL, TR are derived. The light L 0 produced by an incident particle at a distance χ from the left end, will propagate in equal parts towards left and right. Depending on the distance travelled to the P M T s it suffers absorption leading to the signals EL

OC

L 0 · exp( — x/D), E R oc L 0 · exp( — (S — x)/D)

(3)

with D being the absorption coefficient and S the total strip length. The time signals are measured against a reference signal (a fast in-beam start-counter or a high-frequency signal from the accelerator); to simplify matters we assume that this signal precisely denotes the start time of the particles and does not contribute to the time resolution. Therefore the recorded time signals represent the following quantities: TL = TOF + txTR

= TOF + ts_x

(4)

TOF is the time of flight one wants to measure, tx and ts^x are the propagation times of the light to the P M T s . N o t e that the sum tx + ts_x is constant. With the given formulae one finds L0k^El-Er

and

TOF cc j(TL + TR)

(5)

Both quantities are, in first order, independent of the position of incidence. The position itself can be derived in two ways: POS(x)oci(TL-KR)

or

POS{x) oc l n ( £ L / £ R )

(6)

Scintillation Detectors

81

The resolution in TOF is given by (TOF)

= j yjσΐ + σΐ = ο J φ .

( 8 0 % depending on the wavelength), high linearity with the light level, fast time responses (ns) and low noise. The current I (A) generated in a photo diode at the wavelength λ is governed by the p h o t o n flux Ν (λ), the internal photo emission efficiency η(λ), and the reflectivity Ι(λ) =

φ,(λ)Ν(ΧΚ\-Κ(λ))

(7)

Depending on the wavelength of interest, an adapted anti-reflective coating has to be deposited on the detector surface to improve the absorption yield. This can be done by adjusting the refractive index and the thickness of this layer (generally an oxide or a nitride). P h o t o diodes with sensitivity in visible, UV, or even vacuum UV light, are currently available with a very low leakage current, less than 1 nA/cm 2 for a 1 cm 2 surface. A typical photo diode structure is represented in Fig. 6.

B. Lithium-drifted silicon detectors As silicon detectors could be used for spectrometry of β, X-rays and energetic charged particles, the need for wide sensitive layers has made lithium compensation necessary to produce crystals with very low net impurity concentration as long as high resistivity material has not been available from the manufacturers. Nowadays, silicon with net impurity concentration of 1 0 n - 1 0 1 2 / c m 3 is currently fabricated by using advanced refining procedures or neutron compensation. Therefore, lithium compensation remains mainly devoted to the preparation of detectors with sensitive regions several millimetres thick and a nearly uniform electric field for high energy charged particles and very high resolution X-ray spectrometry at liquid nitrogen temperature. An active

Semiconductor Detectors

103

area of more t h a n 10 c m 2 can be p r o d u c e d [55]. As the quality of Si(Li) detectors is now extremely high, the charge collection is nearly perfect a n d resolution limits for X-ray spectrometry become essentially determined by the statistics of pair creation in the material, and by the noise p e r f o r m a n c e of the associated electronics. T h a n k s to the improvements in electronics, resolution of a b o u t 100 eV can be reached for 6 keV X-rays with the aid of pulsed optical or cooled preamplifiers. These detectors can be operated at temperatures slightly lower t h a n L N 2 because the t r a p p i n g levels in Si(Li) are mainly d u e to the primary d o p a n t s Β a n d Li which i n t r o d u c e shallow levels a n d the performances are only deteriorated for t e m p e r a t u r e s below 50 Κ [75].

C. Position sensitive silicon detectors T h e first a t t e m p t s to fabricate position sensitive detectors have m a d e use of surface barriers. M a n y multiple detectors based on individual gold strips with a pitch as low as 100 μιη have been produced [110, 111]. An other concept has been to use a resistive divider by introducing a calibrated resistor on the back contact. T h e pulse height as obtained on this contact is thus position d e p e n d e n t and the energy can be deduced f r o m the signal recorded on the front contact. Spatial resolution better t h a n 1 m m has been d e m o n s t r a t e d [112], An improvement of this concept has been achieved by using implanted contacts with precisely controlled resistive layers [113,114], A position resolution of a fraction of a millimetre with a linearity of a b o u t one percent has been obtained. F u r t h e r m o r e , two dimensional position sensitive counters a n d image capability have been successfully proved by using resistive charge division on b o t h rectifying a n d ohmic contacts [ 113]. Si(Li) position sensitive detectors have also been prepared in a similar way by using the lithium contact as the resistive divider [114]. T h e m a i n drawback of this structure is related to the position dependent R C p r o d u c t . T h e required a d a p t a t i o n of the time constants of the associated electronics limits the achieveable c o u n t rate [115]. A spectacular i m p r o v e m e n t in position sensitive detectors has resulted f r o m the use of existing microelectronics devices on the one h a n d (CCDs), a n d f r o m the specific application of the microelectronics processes to the fabrication of finely p a t t e r n e d detector structures on the other h a n d (especially microstrip detectors a n d silicon drift chambers). T h e i m p o r t a n c e of such devices for high energy physics applications is growing continuously [ 116,117] as well as the extensive invention a n d i m p r o v e m e n t of devices based on equivalent principles. The possibility of using charge coupled devices (CCDs) for bidimensional high energy particle detection was suggested years ago [118,119]. High spatial resolution of 4.3 and 6.1 μιη in o r t h o g o n a l directions was effectively d e m o n s t r a t e d a few years later [120]. A C C D is m a d e of a matrix of pixels connected to r e a d o u t electrodes. T h e schematic principle of one matrix row is shown in Fig. 7. M O S capacitors deposited over the thin space charge layer on a n + — p — p + structure generate a series of potential wells which store the electric charges created by incoming radiation. T h e charges are moved t o w a r d s the r e a d o u t electrode by conveniently changing the potential so that a series readout of the stored image can be p e r f o r m e d . M o s t of the

104

Jean-Pierre P o n p o n One pixel

Fig. 7. Basic Prinicple of the C C D detector

commercially available C C D s are low level optical MOS-type devices realized for imaging on low resistivity and thin (20-30 μπι) silicon epitaxial layers not fully depleted. The resulting number of electron-hole pairs produced by a minimum ionizing particle is therefore very low (about 1500 for a 30 μηι thick sensitive layer). Higher resistivities (1000-2000 Q.cm) and thicker layers (50-100 μπι) allow the formation of a higher signal without position resolution degradation. Construction of pn C C D s constitutes a mean to get fully depleted detectors [121], Such detectors can also be successfully used for X-ray imaging [122]. Typical C C D s present an active area of a few square centimetres with a pixel size down to20-30 μηι. Among all the position sensitive detectors C C D s have the great advantage of being industrially developed devices with applications in many other fields so that the readout electronics are well suited to existing technologies. The long (milliseconds) readout time constitutes the main drawback of these devices. Since the excellent performance of implanted silicon microstrip detectors [123] for minimum ionizing particles detection has been demonstrated [124], extensive attention has been directed towards microstrip and double-sided microstrip vertex detectors for experiments performed by high energy particle accelerators like e + — e~ colliders ( A L E P H [125], D E L P H I [126] collaborations at L E P for example, or C L E O experiment at CESR [127]). The structure of the microstrip detector is basically rather simple, even for double-sided devices with orthogonal strips (Fig. 8). The rectifying side is made of a series of implanted ρ + juncitons of 10 to 100 μηι width with a 20 to 1100 μπι pitch. On the back side the ohmic n + strips have to be separated by a channel stop or a conveniently biased M O S structure to avoid short-circuit due to the influence of the positive oxide charge which induces a near surface accumulation region in the n-type semiconductor. The strip's length can be 20-30 mm and detectors of several cen-

Semiconductor Detectors

105

Oxide Fig. 8. Two-sided microstrip silicon detector

timetres square on large diameter wafers can be built [128]. Charge division between adjacent strips makes the spatial precision better than the pitch so that precisions of 5 - 6 μπι for both sides of double-sided microstrip detectors are obtained for high energy particles. Even for particles impinging the detector at a high angle of incidence (up to 75 degrees) can a good spatial resolution (40 μπι) be maintained [129]. As the time response is roughly the same as that of a simple detector (nanoseconds), high detection rates can be expected. The main complication of such devices comes from readout as connecting each strip to its own preamplifier can no longer be managed if the number of strips becomes too large. One of the present challenges for microstrip detectors, therefore, is concerned with the reduction of the readout connections without too much complication of the fabrication process. Capacitive coupling can serve to reduce the number of connections significantly. This implies that an insulating layer and a second metallization level are deposited over the strips. In additon, biasing resistors can also be integrated on the devices by using polysilicon [128-130], Charge division [131] between 2 or 3 floating strips adjacent to a biased one can also be used to limit the number of readout electrodes. A more precise control of the electrical properties of the strips (impedance, inter strip capacitance) and low noise electronics are then necessary [124], Other approaches combining both resistive charge division and microstrip structure have been proposed to reduce the readout number of electrodes [132], The silicon drift chamber (SDC) is a more sophisticated device based on the principle of the gas drift chamber [133], The sensitive volume is fully depleted from the n + contact and an increasing potential is applied on pairs of opposite ρ + strips so that electrons created by an incoming radiation are drifted towards the n + collecting electrode (Fig. 9). Both the position and the energy deposited can be obtained from the drift time and collected charge [134]. In addition, bidimensional operation is feasible. Detectors of more than 10 cm 2 have been tested successfully with high energy particles [135]. Although the fabrication is more complicated than that of microstrip detectors the S D C present the advantage of bidimensional simpler readout. In addition, their structure leads to every small capacitance and thus low associated noise. The time response is of the order of microseconds.

106

Jean-Pierre Ponpon p+ strips To amplifier

η Si

p+ strips

Fig. 9. Silicon drift chamber (SDC)

Integration on the same silicon substrate of the detector and of some electronics represents the future challenge for monolithic smart pixel detectors. Building small size (μηι χ μπι) detectors together with VLSI circuits on high resistivity substrates needs some adaptation of the microelectronics processes [136, 137] but shows promising results [138], Transistor-based pixel detectors involving new detection principles with amplification and storage capability are presently considered: M O S or J F E T transistors [139, 140], floating base transistor [141]. This approach raises the problem of radiation damage to the integrated electronics circuits which have to undergo the same irradiations as the detectors. Solutions involving the implementation of the electronics on buried oxide layers (SOI structures) could be ways to increase resistance against radiation damage.

D. Radiation damage in silicon detectors A survey of radiation damage in silicon detectors is out of the scope of this presentation. However, a short summary of the most important features as observed when exposing silicon detectors to nuclear radiations will be proposed. The very point to take into account is concerned with the application and the working mode of the detector. As long as high energy resolution is expected, the material properties must remain very good. This means that the minority carriers lifetime which determines the leakage current and therefore the resolution must remain as high as possible. As irradiation has the effect of reducing this by introducing recombination centres in the bandgap the detector properties are very sensitive to rather low irradiation doses. Starting with an initial value τ 0 the lifetime τ depends on the irradiation dose Φ as follows [142, 143]: τ=

τ0/(\+Κττ0Φ)

(8)

Semiconductor Detectors

107

The irradiation constant Κτ is related to the type of radiation. For 1 MeV fast neutrons for example, Κ τ is about 10" 7 cm 2 /s [144] and τ 0 values of no more than a hundred μβ can result from a fairly low fluence. The increase of the diode generation current / g consecutive to a degradation of τ is expressed by: Μ

=qN{W Κ τ Φ

(9)

where N t is the intrinsic carrier concentration (N, = 1.05 10 10 /cm 3 at 300 Κ in silicon). One can notice that for oxide passivated detectors the degradation of the oxide layer under irradiation increases also the surface recombination velocity. As the surface current is generally much lower than the generation current the subsequent fraction of the total leakage current due to surface effects remains however negligible. Irradiation of a silicon detector—which typically presents a minority carrier lifetime of milliseconds—with no more than 10 10 /cm 2 fast neutrons is therefore large enough to deteriorate the spectrometry performances. The acceptable fluence is somewhat higher for light ionizing particles like electrons, close to 10 12 /cm 2 . On the other hand, when the energy resolution is not the parameter of interest and the detector has to work more like a counter, the degradation of the bulk is less critical and much higher fluences can be accommodated. In this case, fluence can become so high that the resistivity of the semiconductor changes due to the introduction of a large concentration of electrically active acceptor defects. Then compensation of the n-type base silicon substrate occurs. Again the actual resistivity can be expressed as a function of the fluence and the initial resistivity: ρ=

ρ0/(1-Κ^μρ0Φ)

(10)

where K(, represents the irradiation constant for change in resistivity. As an example, the resistivity of 2000Q.cm η-type Si should be doubled after irradiation with l.S lO'Vcm 2 fast neutrons. The influence of very high doses (up to 5-10 13 /cm 2 ) of fast neutrons is presently of great concern for the future use of silicon detectors in the next generation of super colliders. A nearly linear increase of the leakage current versus the neutron flux has been observed, in agreement with relation 9. Compensation of η-type silicon (related to a hole trap) has alos been evidenced. Annealing of most of the irradiation defects is induced by heating for a few tens of minutes at a temperature which depends on the involved defects: room temperature (boron-vacancy complex), 100-200°C (phosphorous-vacancy complex), 200-300°C (vacancy-vacancy complex). Partial recovery of the characteristics can be obtained at room temperature by waiting for longer time [145]. Type inversion due to phosphorous deactivation and acceptor creation has also been observed for proton irradiation [95], As a consequence of material and diode degradation, the response of the detector to ionizing particles shows a charge collection defect. However, the pulse response remains essentially the same in terms of time duration [146], so that the behaviour as a particle counter can be still useful. Especially, one can notice that in spite of bulk degradation the performances of microstrip detectors under severe irradiation may remain acceptable [147],

108

Jean-Pierre Ponpon

Ε. Amorphous silicon detectors In view of preparing large area systems for position detection the fabrication of silicon devices in the a m o r p h o u s phase has been given some attention [148-150]. Amorphous silicon is a semiconductor which when hydrogenated presents interesting electronic properties. The b a n d g a p of a-Si: Η is somewhat higher than that of crystalline silicon and the average energy for pair creatin is 4.3 eV. Due to the structure, the mobility is very low, 2 cm 2 /V-s for electrons and 5 1 0 " 3 cm 2 /v-s for holes, respectively. Nevertheless, for thin layers (less than 50 μηι) p - i - n diodes made on amorphous silicon present good time response (nanoseconds) and are able to detect minimum ionizing particles with a signal/noise ratio high enough. Improvement of the detectors structure to increase the charge collection could lead to the building of cheap large area position detectors more durable against radiation than conventional crystalline Si devices.

V. Exotic materials A common feature of all the compound semiconductor materials which may be used instead of germanium and silicon is concerned with crystal preparation. Purification of the base elements, control of the residual impurities and of the stoichiometry during growth, and fabrication of high quality and high volume single crystals free of defects constitute the source for most of the limitations of these detectors. Improvement of these parameters can be considered as the main challenge for future developments and continuous effort in this way is being made, including micro gravity growth in space [151,152]. Although many semiconductors have shown the possibility of being used for detection (GaAs, InP, CdSe, GaSe, P b l 2 , . . . ) most of the work in this field has been dealing with cadmium telluride and more recently with mercuric iodide which can now be manufactured on a commercial scale.

A. Cadmium telluride Cadmium telluride detectors find a wide field of application as γ-ray spectrometers or counters for dosimetry, nuclear medicine or process control instrumentation [153], An energy resolution of 1 keV for 6 keV X-rays (Fe ka) and better than 3 keV for 122 keV γ-rays ( 5 7 Co) is recorded for detectors with a sensitive volume of one cubic centimetre [154]. High counting rates (10 5 /s) can be accommodated by CdTe detectors. Even higher rates may be envisaged by using a current working mode instead of a pulse working mode. Depending on the growth method and compensation use, CdTe can be prepared with resistivity up to about 109 Q.cm. The room temperature performance of detectors built on such material can become comparable to results obtained with Ge at 77K for γ-spectrometry. The presence of residual impurities, of compensation centres and of stoichiometric deviations induce trapping levels which rapidly degrade the detectors performances when increasing the sensitive volume. In addition, in any crystals these defects may be the source for a time dependent response to irradiation, resulting from a modification of the electric field inside the depleted zone (polarization

Semiconductor Detectors

109

effect). Loss of the counting efficiency and degradation of the energy resolution result from this behaviour.

B. Mercuric iodide As a high atomic number material with a large bandgap, mercuric iodide is a good candidate for room temperature operation with very low noise and high photoelectric efficiency. Large single crystals of a few cubic centimetres can be manufactured but the low carrier mobilities, especially that of holes, limit the usable thickness to the range of millimetres as far as spectrometry is expected. Consequently, Hgl 2 detectors are mainly used for low energy γ-rays and X-rays although detection of conversion electrons can also be performed. Single detectors with a sensitive surface in the range of mm 2 to cm 2 working at room temperature present an energy resolution better than 200 eV, depending on the electronics, for X-rays in the range 0.85-5.9 keV [154-156], This must be compared to the 100 eV energy resolution as obtained with the best Si(Li) detectors cooled at liquid nitrogen temperature. The high counting rate capability (up to 2-105/s) makes these detectors interesting for synchrotron radiation monitoring [155], On the other hand, X and γ imaging up to 662 keV with a spatial resolution better than 2 mm has been demonstrated [157] and improvements in this direction can be expected. Hgl 2 devices can also be used as photo diodes or counters for dosimetry [154]. One can notice that mercuric iodide is a very reactive material and long term stability requires an efficient coating to be deposited over the detectors.

VI. Conclusion The recent evolution of the detector field has been marked by a more accurate control of the fabrication processes, both for base materials growth and for the preparation of devices. This has led to the possibility of manufacturing low noise, large are acounters. At present, 70-80 mm diameter Ge and Si crystals of detector grade quality are being processed. Among the new applications of these conventional detectors one can notice the growing interest for high energy particles identification, by simultaneous measurement of both ionization and phonons at very low temperature ( i n the range of a few tens of mK). On the other hand, the revolutionary advent of the planar process has made new types of silicon detectors feasible. The first consequence of these advances has been a significant improvement of the properties of the conventional-type detectors. However, the most significant feature is concerned with the possibility of building large systems involving a great number of elementary detectors either for high efficiency γ-ray sepctrometry (association of tens of large volume Ge detectors), or for track reconstitution with spatial resolution in the micrometer range (large area of silicon position detectors). Fabrication of systems including detectors and VLSI electronics (smart pixel detectors) will constitute the next step to overcome the increasing complexity of such multidetector arrays. However, the question of the choice between industrially developed devices (CCDs) and sophisticated devices

110

Jean-Pierre Ponpon

specifically built for unique application in high energy physics (microstrips, SDC, pixel detectors) is still an open one. Amorphous silicon detectors could also present interesting possibilities in this area. The case of exotic materials is somewhat different as further improvement of their performances remains strongly dependent on the increase of the crystals quality and volume. Nevertheless, for applications needing room temperature low energy γ- and X-ray spectrometery and imaging with small devices, they can already favourably compete with Si and Ge. The possibility of using semiconductor detectors under strong irradiation will make it necessary to develop the detector technology of other materials such as C or SiC which is still in its early stages.

References [1] McKay, K. G„ A Germanium Counter, Phys. Rev. 176, 1537, 1949. [2] Debiak, T. W„ Hughlock, B. W., Bocskor, S. J., D'Agostino, M. D„ Scheid, Ε. J., Kesselman, M., Misita, R. M., Detection of Gamma Rays Using a Coupled Array of High-Purity Germanium Detectors, IEEE Trans. Nucl. Sei. NS 37, n6, 1659, 1990. [3] Fisher, T. R., Hamilton, J. W„ Hawley, J. D„ Kilner, J. R., Murphy, M. J., Nakano, G. H„ Imaging of Gamma-Rays with the WINKLER High-Resolution Germanium Spectrometer, IEEE Trans. Nucl. Sei. Ns 37, n3, 1483, 1990. [4] Beausang, C. W., et. al., Measurements on Prototype Ge and BGO Detectors for the Eurogam Array, Nucl. Instr. Meth. A313, 37, 1992. [5] Von Ammon, W. and Herzer, H., The Production and Availability of High Resistivity Silicon for Detector Application, Nucl. Instr. Meth. 226,94,1984. [6] Martoff, C. J., Kaczatovitcz, E., Neubauer, Β. J., Lopez, E., Zhang, Y., Ziemba, F. P., Operation of a High Purity Silicon Diode Alpha Particle Detector at 1.4 K, Nucl. Instr. Meth. A301, 376, 1991. [7] Freck, D. V., and Wakefield, J., Gamma Ray Spectrum Obtained with a Lithium-Drifted p - i - n Junction in Germanium, Nature 1193, 669, 1962. [8] Webb, P. P., and Williams, R. L., Gamma Ray Spectroscopy Using a Germanium Lithium-Drifted Diode, Nucl. Instr. Meth. 22, 361, 1963. [9] Tavendale, A. J., and Ewan, G. T., A High Resolution Lithium-Drifted Germanium Gamma Ray Spectrometer, Nucl. Instr. Meth. 25, 185, 1963. [10] Pell, Ε. M„ Ion Drift in an n - p Junction, J. Appl. Phys. 31, 291,1960. [11] Grahmann, H. and Kalbitzer, S., Ge(Li) Detectors with Highly Compensated Intrinsic Regions, Nucl. Instr. Meth. 101, 61,1972. [12] Coche, Α., Siffert, P., Cappellani, F., Resteiii, G., in Semiconductor Detectors, Bertolini and Coche, North Holland 1968, chap. 2. [13] Henck, R., Stuck, R., Siffert, P., Schoenmakers, W., DeLaet, L. H., Influence of Selected Chemical Impurities on the Trapping Parameters of Ge(Li) Counters, IEEE Trans. Nucl. Sei. NS 19, n3, 299, 1972. [14] Henck, R., Siffert, P., Schoenmakers, W., DeLaet, L. H., Trap Parameters of Ag, Co, Cr and Te Impurities in Ge(Li) Detectors, Nucl. Instr. Meth. 10, 61, 1972. [15] Schoenmakers, W., Van Ouytsel, J. I., DeLaet, L. H., PCD Lifetime Measurements as a Function of Temperature in Detector Grade and Impurity Doped Germanium, IEEE Trans. Nucl. Sei. NS 19, nl, 279, 1972. [16] McMath, Τ. Α., and Sakai, E., A Search for Deep Trap Levels in Ge(Li) Detectors, IEEE Trans. Nucl. Sei. NS19, nl, 289,1972. [17] Stuck, R„ Ph. D. Thesis, University of Strasbourg, 1976. [18] Malm, H. L., Tavendale A. J., and Fowler, I. L., A Large Volume, Coaxial, Lithium-Drift, Germanium Gamma-Ray Spectrometer, Can. J. Phys. 43, 1173, 1965.

Semiconductor Detectors

111

Malm, H. L., Improvement of Large Volume Coaxial Germanium Spectrometers, IEEE Trans. Nucl. Sei. NS 13, n3, 285, 1966. Hotz, Η. P., Mathiesen, J. M., Hurley, J. P., Calculated Efficiencies of Ge(Li) Detectors, Nucl. Instr. Meth. 37, 93, 1965. Griffiths, R., Calculated Total Efficiencies of Coaxial Ge(Li) Detectors, Nucl. Instr. Meth. 91, 377, 1970. Lai, B., Iyengar, Κ. V., Monte Carlo Calculations of Gamma-Ray Response Characteristics of Cylindrical Ge(Li) Detectors, Nucl. Instr. Meth. 79, 19, 1970. Seyfarth, H., Hassan, A. M., Hrastnik, B., Gottel, P., Delang, W., Efficiency Determination for Some Standard Type Ge(Li) Detectors for Gamma-Rays in the range from 0.04 to 11 MeV, Nucl. Instr. Meth. 105, 301, 1972. Riepe, G., Protic, D., Reich, J., High Resolution Charged Particles Spectrometry with Side-Entry Ge(Li) Detectors in the Energy Range of the Jülich Isochronous Cyclotron, Nucl. Instr. Meth. 124, 527, 1975. Ryvkin, S. M., Matveev, Ο. Α., Strokan, Ν. B., Khusainov, A. Kh., A Gamma-Ray Spectrometry Counter Based on Germanium with Radiation Defects, Sov. Phys. Doklady 10, n i l , 1116, 1966. Llacer, J., Basic Limitations in the Fabrication of Radiation Detectors by G a m m a Compensation of Germanium, IEEE Trans. Nucl. Sei. NS 19, nl, 295, 1972. Hall, R. N., Baertsch, R. D., Soltys, T. J., Petrucco, L., High Purity Germanium for Gamma Detectors, General Electric annual reports 1 to 4, NYO- 3870, 1968-1971. Baertsch, R. D. and Hall, R. N., Gamma-Ray Detectors Made from High Purity Germanium, IEEE Trans. Nucl. Sei. NS 17, n3, 235, 1970. Hall, R. N., and Soltys, T. J., High Purity Germanium for Detector Fabrication, IEEE Trans. Nucl. Sei. NS 18, η 1, 160,1971. DeLaet, L. H., Schoenmakers, W. K., Guislain, H. J., Meeus, M., Quality Control of p-Type and Intrinsic Ge Monocrystals for γ-Radiation Detectors, Nucl. Instr. Meth. 101, 11,1972. Hansen, L. H., High Purity Germanium Crystal, Growing Nucl. Instr. Meth. 94, 377, 1971. Haller, Ε. E., Hansen, W. L., Goulding, F. S., Distribution of Donors and Acceptors in High Purity Germanium Crystals, IEEE Trans. Nucl. Sei. NS 20, nl, 481, 1973. Hall, R. N., Chemical Impurities and Lattice Defects in High Purity Germanium, IEEE Trans. Nucl. Sei. NS 21, nl, 260, 1974. Haller, Ε. E., Hansen, W. L., Impurities in High Purity Germanium as Determined by Fourier Transform Spectroscopy, IEEE Trans. Nucl. Sei. NS 21, nl, 279, 1974. Hansen, W. L., Haller, Ε. E., A View of the Present Status and Future Prospects of High Purity Germanium, IEEE Trans. Nucl. Sei. NS 21, nl, 251, 1974. Haller, Ε. E., Li, P. P., Hubbard, G. S., Hansen, W. L., Deep Level Transient Spectroscopy of High Purity Germanium Diodes/Detectors, IEEE Trans. Nucl. Sei. NS 26, n l , 265, 1979. Evwaraye, A. O., Hall, R. N., Soltys, T. J., DLTS Measurements of Trapping Defects in High Purity Germanium, IEEE Trans. Nucl. Sei. NS 26, nl, 297, 1979. Schoenmaekers, W. K., Clauws, P., Van den Steen, K., Broeckx, J., Henck, R., Carrier Trapping in High Purity Germanium, IEEE Trans. Nucl. Sei. NS 26, nl, 256, 1979. Simoen, E., Clauws, P., Broecks, J., Vennik, J., Van Sande, Μ., DeLaet, L., Correlation Between DLTS Measurements and the Performance of High Purity Germanium Detectors, IEEE Trans. Nucl. Sei. NS 29, nl, 789, 1982. Hall, R. N., Racette, J. H., Diffusion and Solubility of Copper in Extrinsic and Intrinsic Germanium, Silicon and Gallium Arsenide, J. Appl. Phys. 35, 379, 1964. Ponpon, J. P., Grob, J. J., Stuck, R., Burger, P. Siffert, P., Boron Implanted Contacts on High Purity Germanium, Proc. II int. Conf. on Ion Implantation, Garmisch-Partenkirchen, 1971,1. Ruge and J. Graul eds., Springer Verlag, ρ 420. Herzer, Η., Kalbitzer, S., Ponpon, J. P., Stuck, R., Siffert, P., Ion Implanted High Purity Germanium Detectors, Nucl. Instr. Meth. 101, 31, 1972.

112

Jean-Pierre Ponpon

[43] Pehl, R. H., Cordi, R. C., Goulding, F. S., High-Purity Germanium: Detector Fabrication and performance, IEEE Trans. Nucl. Sei. NS 19, nl, 265, 1972. [44] Hubbard, G. S., Haller, Ε. E., Hansen, W. L., Ion Implanted η-Type Contact for High Purity Germanium Radiation Detectors, IEEE Trans. Nucl. Sei. NS 24, nl, 161, 1977. [45] Riepe, G., Protic, D., High Purity Germanium Detectors with Both Contacts made by Ion Implantation, Nucl. Instr. Meth. 165, 31, 1979. [46] Dinger, R. J., Dead Layers at the Surface of p - i - n Detectors: A Review, IEEE Trans. Nucl. Sei. NS 22, η 1, 135, 1975. [47] Hansen, W. L., Haller, Ε. E., Hubbard, G. S., Protective Surface Coatings of Semiconductor Nuclear Radiation Detectors, IEEE Trans. Nucl. Sei. NS 27, nl, 247,1980. [48] Allsworth, F. L., Howes, J. H., White, G., Intrinsic Germanium Detectors used in a High Counting Rate System, Nucl. Instr. Meth. 193, 57, 1982. [49] Baertsch, R. D., Surface Effects on p-Type High Purity Germanium Detectors at 77K, IEEE Trans. Nucl. Sei. NS 21, nl, 347, 1974. [50] Gelezunas, V. L., The Effect of Exposure to Various Environments on the Subsequent Performance of High Purity Germanium Gamma-Ray Detectors, IEEE Trans. Nucl. Sei. NS 21, n l , 360, 1974. [51] Malm, H. L., Dinger, R. J., Charge Collection in Surface Channels on High Purity Germanium Detectors, IEEE Trans. Nucl. Sei. NS 23, nl, 76, 1976. [52] Llacer, J., Planar and Coaxial High Purity Germanium Detectors, Nucl. Instr. Meth. 98, 259, 1972. [53] Marler, J. M., Hewka, P. V., Coaxial Detectors from H. P. Germanium, IEEE Trans. Nucl. Sei. NS 21, n l , 287, 1974. [54] Keyser, R. M., Raudorf, T. W., Germanium Radiation Detectors manufacturing: Process and Advances, Nucl. Instr. Meth. A286, 357,1990. [55] Henck, R., Intertechnique Capabilities in Detector Fabrication, Nucl. Instr. Meth. A228, 278, 1990. [56] Verplancke, J., Burger, P., Schoenmaekers, W., Canberra Semiconductors, an Industrial Partner for Physics Research, Nucl. Instr. Meth. A288, 287, 1990. [57] Brodzinski, R. L., et al., An Ultralow Background Germanium Gamma Ray Spectrometer, Nucl. Instr. Meth. A239, 207, 1985. [58] Protic, D., Riepe, G., Position Sensitive Germanium Detectors, IEEE Trans. Nucl. Sei. NS 32,nl, 553,1985. [59] Luke, P. N., Madden, N. W., Goulding, F. S., Germanium Detectors with a Built-in Transverse Drift Field, IEEE Trans. Nucl. Sei. NS 32, nl, 457, 1985. [60] Gutknecht, D., Photomask Technique for Fabricating High Purity Germanium Strip Detectors, Nucl. Instr. Meth. A288, 13, 1990. [61] Amendolia, S. R. et al., Germanium Microstrip Detectors with 50 and 100 μπι Pitch, Nucl. Instr. Meth. 226, 117, 1984. [62] Poenaru, D. N., Collection Time of Electron-Hole Pairs in a Coaxial Ge(Li) Radiation Detector, IEEE Trans. Nucl. Sei. NS 14, n5,1,1967. [63] Sakai, E., Charge Collection in Coaxial Ge(Li) Detectors, IEEE Trans. Nucl. Sei. NS 15, n3, 310,1968. [64] Poenaru, D. N., Stuck, R., Siffert, P., Collection Efficiency and Charge Carrier Losses in Coaxial and Planar Ge(Li) Detectors. Influence on the Time Resolution, IEEE Trans. Nucl. Sei. NS 17, n3, 176, 1970. [65] Strauss, M. G., Larsen, R. N., Sifter, L. L., Pulse Shape Distributions from Gamma Rays in Lithium Drifted Germanium Detectors, IEEE Trans. Nucl. Sei. NS 13, n3, 265, 1966. [66] Matsushita, N., McHarris, W. C., Firestone, R. B., Kasagi, J., Kelly, W. H., On Improving Ge Detector Energy Resolution and Peak upon Compton Ratios by Pulse Shape Discrimination, Nucl. Instr. Meth. 179, 119, 1981. [67] Moszynski, M., Duchene, G., Ballistic Deficit Correction Methods for Large Ge Detectors, Nucl. Instr. Meth. A308, 557, 1991.

Semiconductor Detectors

113

[68] Raudorf, Τ. Μ., Paulus, J., Bedwell, Μ. Ο., Martini, Μ., Comparative Timing Performance of Large Volume Ge(Li) and H P G e Coaxial Detectors, IEEE Trans. Nucl. Sei. NS 24, η 1,78, 1977. [69] Kraner, Η. W., Chasman, C., Jones, K. W., Effects produced by Fast Neutron Bombardment of Ge(Li) Gamma Ray Detectors, Nucl. Instr. Meth. 62, 173, 1968. [70] Llacer, J., Kraner, H. W., Neutron Damage and Annealing in High Purity Ge Radiation Detectors, Brookhaven National Laboratory Report BNL-16142, 1971. [71] Goulding, F. S., Pehl, R. H., A Survey of Radiation Damage in Semiconductor Detectors, IEEE Trans. Nucl. Sei. NS 19, nl, 91, 1972. [72] Darken, L. S., Trammel, R. C., Raudorf, T. W., Pehl, R. H., Neutron Damage In Ge(HP) Coaxial Detectors, IEEE Trans. Nucl. Sei. NS 28, nl, 572, 1981. [73] Kraner, H. W., Pehl, R. H., Haller, Ε. E., Fast Neutron Radiation Damage of High Purity Germanium Detectors, IEEE Trans. Nucl. Sei. NS 22, n l , 149, 1975. [74] Pehl, R. H., Varnell, L. S., Metzger, A. E., High Energy Proton Radiation Damage of High Purity Germanium Detectors, IEEE Trans. Nucl. Sei. NS 25, nl, 409, 1978. [75] Martini, M., McMath, Τ. Α., Trapping and Detrapping Effects in Lithium-Drifted Germanium and Silicon Detectors, Nucl. Instr. Meth. 79, 259, 1969. [76] Zanio, K. R., Akutagawa, W. M., Charge Transport in Semiconductor Detectors with Trapping and Detrapping, IEEE Trans. Nucl. Sei. NS 15, n3, 260, 1968. [77] Stuck, R., Ponpon, J. P., Siffert, P., Ricaud, C., Behavior of High Purity Germanium Nuclear Counters at Low Temperature, Appl. Phys. Lett. 19,461, 1971. [78] Pehl, R. H., Haller, E. E., Cordi, R. C., Operational Characteristics of Germanium Detectors at Higher Temperatures, IEEE Trans. Nucl. Sei. NS 20, n l , 494, 1973. [79] Nakano, G. H., Simpson, D. Α., Imhof, W. L., Characteristics of Large Intrinsic Germanium Detectors Operated at Elevated Temperatures, IEEE Trans. Nucl. Sei. NS 24, n l , 68, 1977. [80] Mayer, J. W., Gossick, B., Use of Gold Germanium Broad Area Barriers as Alpha Particle Spectrometers, Rev. Sei. Instr. 27, 407,1956. [81] Mayer, J. W., Performance of Germanium and Silicon Surface Barrier Diodes as Alpha Particles Spectrometers, J. Appl. Phys. 30, 1937, 1959. [82] Walter, F. J., Dabbs, J. W., Roberts, L. O., Large Area Germanium Surface Barrier Counters, Rev. Sei. Instr. 31, 756, 1960. [83] Blankenship, J. L., Borkowski, C. J., Silicon Surface Barrier Nuclear Particle Spectrometer, IRE Trans. Nucl. Sei. NS 7, n2-3, 190,1960. [84] Dearnaley, G., Whitehead, A. B., How Harwell Makes Surface Barrier Detectors, Nucleonics 19, 72, 1961. [85] Inskeep, C., Elad, E., Sareen, R. Α., Surface Barrier Structures with Various Metal Electrodes, IEEE Trans. Nucl. Sei. NS 21, 374, 1974. [86] Gibbons, P. E., Metal-Semiconductor-Contacts for Solid State Radiation Detectors, Nucl. Instr. Meth. 29, 289, 1964. [87] Siffert, P., Laustriat, G., Coche, Α., Rectifying Process in Surface Barrier Detectors, IEEE Trans. Nucl. Sei. NS 11, n3, 244, 1964. [88] Hansen, N. J., Scott, R. G., Henderson, D. J., A Position Sensitive Surface Barrier Array Type Detector, Nucl. Instr. Meth. 105, 293, 1972. [89] Anderson, L. P., Berg, S., Mechanism of Aluminum Contacts in Surface Barrier Detectors, Nucl. Instr. Meth. 108,435, 1973. [90] England, J. B., Hammer, V. W., A New Type Of Non-Injecting, Back Contact for totally depleted Surface Barrier Detectors, Nucl. Instr. Meth. 96, 81, 1971. [91] Siffert, P., Coche, Α., New Results Concerning the Rectifying Process in Surface Barrier Counters, IEEE Trans. Nucl. Sei. NS 12, nl, 284, 1965. [92] Adeichickov, V. V., Some Properties of Si(p) Detectors, Nucl. Instr. Meth. 155, 125, 1978. [93] Shiraishi, F., Hosoe, M., Takami, Y., Ohsawa, Y., Totally Depleted Surface Barrier Detectors Made of Ultra-High Purity p-Type Silicon Crystals, IEEE Trans. Nucl. Sei. NS 29, nl,775, 1982.

114

Jean-Pierre Ponpon

[94] Fox, R. J., Borkowski, C. J., Silicon Surface Barrier Detectors with High Reverse Breakdown Voltage, IRE Trans. Nucl. Sei. NS 9, 213, 1962. [95] Fretwurst, E., Grube, R., Lindström, G., Nagel, J., Development of Large Area Silicon Detectors, Nucl. Instr. Meth. A253,467, 1987. [96] Fretwurst, E., Herdan, Η., Lindström, G., Pein, V., Rollwagen, Μ., Schutz, Η., Thomsen, P., Wunstorf, R., Silicon Detector Developments for Calorimetry: Technology and Radiation Damage, Nucl. Instr. Meth. A288, 375, 1990. [97] Goulding, F. S., Hansen, W. L., Leakage Current in Semiconductor Junction Radiation Detectors and its Influence on Energy-Resolution Characteristics, Nucl. Instr. Meth. 12, 149, 1961. [98] Kalbitzer, S., Bader, R., Herzer, H., Bethge, K., The Fabrication of High Quality Silicon Junction Detectors by Low Energy Ion Implantation, Z. Physik 203, 117, 1967. [99] Meyer, O., Haushahn, G., High Resolution Particle Detectors Produced by Ion Implantation, Nucl. Instr. Meth. 56,177, 1967. [100] Keil, G., Lindner, E., Low-Noise Oxide Passivated p + — η Silicon Detectors, Nucl. Instr. Meth. 101, 43, 1972. [101] Kemmer, J., Fabrication of Low Noise Silicon Radiation Detectors by the Planar Process, Nucl. Instr. Meth. 169, 499, 1980. [102] Inskeep, C., Elad, Ε. E., Sareen, R. Α., Surface Barrier Structures with Various Metal Electrodes, IEEE Trans. Nucl. Sei. NS 21, 379,1973. [103] Rossington, C. S., Walton, J. T., Jacklevic, J. M., Si(Li) Detectors with Thin Dead Layers for Low Energy X-Ray Detection, IEEE Trans. Nucl. Sei. NS 38, n2, 239, 1991. [104] Grahmann, H. G., Kalbitzer, S., Measurement of Ultra-Thin Windows of Ion Implanted Silicon Detectors with Low Energy Proton Beams, Nucl. Instr. Meth. 136, 145, 1976. [105] Wilkins, B. D„ Fluss, Μ. J., Kaufman, S. B., Gros, C. E , Steinberg, P., Pulse Height Defect for Heavy Ions in a Silicon Surface Barrier Detector, Nucl. Instr. Meth. 92, 381, 1971. [106] Ponpon, J. P., Siffert, P., Vazeille, F., Thin dE/dx Detectors of Uniform Thickness Made on Epitaxial Silicon, Nucl. Instr. Meth. 112, 465, 1973. [107] Stab, L., Thin Epitaxial Silicon Detectors, Nucl. Instr. Meth. A288, 24, 1990. [108] Tuzzolino, A. J., Long Time Photo and Charged Particle Response of Silicon Surface Barriers, Nucl. Instr. Meth. 204,161,1982. [109] Blanar, G., et al., Photodiodes Readout for Scintillating Crystals of BGO and Nal(Tl), Nucl. Instr. Meth. 203, 213, 1982. [110] Dearnaley, G., Whitehead, A. B., The Semiconductor Surface Barrier for Nuclear Particle Detection, Nucl. Instr. Meth. 12, 205,1961. [111] Hansen, N. J., Henderson, D. J., Scott, R. G., Surface Barrier Particle Detectors, Nucl. Instr. Meth. 104, 333, 1972. [112] Bock, R., Duhm, Η. H„ Melzer, W., Puhlhofer, F., Stadler, B„ An Array of Position Sensitive Surface Barrier Detectors for Use in a Broad Range Magnetic Spectrograph, Nucl. Instr. Meth. 41, 190, 1966. [113] Kalbitzer, S., Bader, R., Melzer, W., Stumpft, W., An X - Y Position Sensitive Detector, Nucl. Instr. Meth. 54, 323, 1967. [114] Laegsgaard, E., Martin, F. W., Gibson, W. M., Position Sensitive Detectors Made by Ion Implantation in Silicon, Nucl. Instr. Meth. 60, 24, 1968. [115] Gigante, J. R., Si(Li) Position Sensitive Detector, Nucl. Instr. Meth. I l l , 345, 1973. [116] Heijne, Ε. H., Jarron, P., Development of Silicon Pixel Detectors: An Introduction, Nucl. Instr. Meth. A275,467, 1989. [117] Karchin, P. E., Use of Pixel Detectors in Elementary Particle Physics, Nucl. Instr. Meth. A305, 497, 1991. [118] Gordon, Ε. I., Charge Couplin; an All Electronic Approach to Position Sensitive particle Detection, IEEE Trans. Nucl. Sei. NS 19, n3, 190, 1972. [119] Damerell, C. J., Farley, F. J., Gillman, A. R., Wickens, F. J., Charge Coupled Devices for Particle Detection with High Spatial Resolution, Nucl. Instr. Meth. 185, 33, 1981.

Semiconductor Detectors

115

120] Bailey, R., Damereil, C. J., English, R. L , Gillman, A. R„ Lintern, A. L., Watts, S. J., Wickens, F. J., First Measurement of Efficiency and Precision of C C D Detectors for High Energy Physics, Nucl. Instr. Meth. 213, 201, 1983. 121] Strüdel, L„ et al., T h e M P I / A I T X-Ray Imager (MAXI) High Speed p - n C C D s for X-Ray Detection, Nucl. Instr. Meth. A288, 227, 1990. 122] Acton, P., et al., F u t u r e Potential of Charge Coupled Devices as Detectors of Ionizing Radiations, Nucl. Instr. Meth. A305, 504, 1991. 123] Kemmer, J., Burger, P., Henck, R., Heijne E., Performance and Application of Passivated Ion Implanted Silicon Detectors, I E E E Trans. Nucl. Sei. N S 29, n l , 733, 1982. 124] Hyams, B., Koetz, U., Belau, Ε., Klanner, R., Lutz, G., Neugebauer, Ε., Wylie, Α., Kemmer, J., A Silicon Counter Telescope to Study Short-Lived particles in High Energy H a d r o n i c Interactions, Nucl. Instr. Meth. 205, 99,1983. 125] Becker, H., et al., New Development in Double-Sided Silicon Strip Detectors, I E E E Trans. Nucl. Sei. N S 37, 101, 1991. 126] Turchetta, R., Measurement of Spatial Resolution of Double-Sided Double-Metal A C Coupled Silicon Microstrip Detectors, Sixth European Symposium on Semiconductor Detectors, Milan, 1992. 127] Alexander, J. P., et al., Studies of Performance and Resolution of Double-Sided Silicon Strip Detectors, I E E E Trans, Nucl. Sei. N S 38, n2, 263, 1991. 128] Hietanen, I., et al., Ion Implanted Capacitively Coupled Silicon Strip Detectors with Integrated Polysilicon Bias Resistors Processed on a 100 m m Wafer, Nucl. Instr. Meth. A310, 671,1991. 129] Hanai, H., et al., Spatial Resolution of Silicon Microstrip Detector for Particle With Large Incident Angle, Nucl. Instr. Meth. A314, 495, 1992. 130] Ten Kate, W. R., Heijne, Ε. H., Capacitive Multiplexing on a Silicon Microstrip Detector, Nucl. Instr. Meth. A234, 398, 1985. 131] Kötz, V., et al., Silicon Strip Detectors with Capacitive Charge Division, Nucl. Instr. Meth. A235, 481, 1985. 132] Barabash, L. S., et al., Semiconductor Microstrip Detector with a Resistive Layer, Nucl. Instr. Meth. A288, 375, 1990. 133] Gatti, E., and Rehak, P., Semiconductor Drift Chamber. An Application of a Novel Charge Transport Scheme, Nucl. Instr. Meth. 225, 608, 1984. 134] Rehak, P., Gatti, E., Longoni, Α., Kemmer, J., Holl, P., Klanner,R., Lutz, G., Wylie, Α., Semiconductor Drift C h a m b e r s for Position and Energy Measurements, Nucl. Instr. Meth. A235, 224, 1985. 135] Cox, T. P., et al., Beam test of a Large Area Semiconductor Drift Detector, Sixth European Symposium on Semiconductor Detectors, Milan, 1992. 136] Vanstraelen, G., Debusschere, I., Claeys, G., Declerk, G., New Concepts for Integrated Solid State Detector Electronics, Nucl. Instr. Meth. A273, 625, 1988. 137] Wouters, S. E., Audet, S. Α., Kim, Μ. H., Schooneveld, Ε. M., Integrating transistors on High O h m i c Silicon, Nucl. Instr. Meth. A288, 44, 1990. 138] Kenney, C., Parker, S., Snoeys, W., Plummer, J., Rosseel, G., Aw, C. H., First Beam Test Results F r o m a Monolithic Silicon Pixel Detector, Sixth E u r o p e a n Symposium on Semiconductor Detectors, Milan, 1992. 139] Kemmer, J., et al., Experimental Confirmation of a New Semiconductor Detector Principle, Nucl. Instr. Meth. A288, 92, 1990. 140] Klein, P., et al., Design and Performance of Semiconductor Detectors with Integrated Amplification and Charge Storage Capability, Nucl. Instr. Meth. A305, 517, 1991. 141] Ardelean, J., Chase, R. L., Hrisoho, Α., A Proposed Pixel Detector Based on Floating Base Transistor, Nucl. Instr. Meth. A305, 549, 1991. 142] Van Lint, A. J., Gigas, G., Barengoltz, J., Correlation of Displacement Effects Produced by Electrons protons and Neutrons in Silicon, I E E E Trans. Nucl. Sei. N S 22, n6, 2663, 1975. 143] Kraner, H. W., Radiation Damage in Semiconductor Detectors, I E E E Trans. Nucl. Sei. N S 29, n3, 1088, 1982.

116

Jean-Pierre Ponpon

[144] Kraner, Η. W., Li, Ζ., Posnecker, Κ. U., Fast Neutron Damage in Silicon Detectors, Nucl. Instr. Meth. A279, 266, 1989. [145] Li, Z., Chen, W., Kraner, H. W., Effects of Fast Neutron Radiation on the Electrical Properties of Silicon Detectors, Nucl. Instr. Meth. A308, 585, 1991. [146] Lemeilleur, F., Glaser, Μ., Heijne, Ε. Η., Jarron, P., Occelli, Ε., Neutron Induced Radiation Damage in Silicon Detectors, 1991 IEEE Nuclear Symposium and Medical Imaging Conference, Santa-Fe, USA. [147] Pitzl, D., et al., Type Inversion in Silicon Detectors, Nucl. Instr. Meth. A311, 98, 1992. [148] Perez-Mendez, V., Morel, J., Kaplan, S. N., Street, R. Α., Detection of Charged Particles in Amorphous Silicon Layers, Nucl. Instr. Meth. A252, 478, 1986. [149] Perez-Mendez, V., Kaplan, S. N., Cho, G., Fujieda, I., Qureshi,S., Ward, W„ Street, R. Α., Hydrogenated Amorphous Silicon Pixel Detectors for Minimum Ionizing Particles, Nucl. Instr. Meth. A273, 127, 1988. [150] Hamel, L. Α., Dubeau, J., Pochet, Τ., Equer, Β., Signal Formation in a —Si:H particle Detectors, IEEE Trans. Nucl. Sei. NS 38, n2, 251, 1991. [151] Van den Berg, L., Schnepple, W. F., Mercuric Iodide Crystal Growth in Space, Nucl. Instr. Meth. A283, 335,1989. [152] Siffert, P., et al., Characterization of CdTe Crystals Grown under Microgravity Conditions, Nucl. Instr. Meth. A289, 363, 1989. [153] Entine, G., Waer, P., Tiernan, T., Squillante, R., Survey of CdTe Nuclear Detector Applications, Nucl. Instr. Meth. A283, 282, 1989. [154] Beyerle, A. G., Hull, K. L., Neutron Detection with Mercuric Iodide Detectors, Nucl. Instr. Meth. A256, 377, 1987. [155] Iwanczyc, J. S., Advances in Mercuric Iodide X-Ray Detectors and Low Noise Preamplification Systems, Nucl. Instr. Meth. A283, 208, 1989. [156] Iwanczyk,J. S., Wang, J. J., Bradley, J. G., Albee, A. L., Schnepple, W. F., Advances in the Developement of Encapsulants for Mercuric Iodide X-Ray Detectors, IEEE Trans. Nucl. Sei. NS 37, n6, 2214, 1990. [157] Patt, Β. E., Beyerle, A. G., Dolin, R. C., Ortale, C., Developments in Mercuric Iodide Gamma Ray Imaging, Nucl. Instr. Meth. A283, 215, 1989.

4

Track Detectors and Soleno applied for Cluster Radioactivities

E. Hourany,

I. H. Plonski and D. N.

Poenaru

Table of contents I. Introduction 117 II. The magnetic spectrometer S O L E N O 119 A. Superconducting solenoidal spectrometer 119 B. Sources 120 C. Detection system 122 D. Experiments of cluster radioactivity and its fine structure III. Solid state nuclear track detectors 125 A. Track detector characteristics and materials 125 B. Background and the loss of sensitivity 127 C. Irradiation geometry and microscope scanning 128 D. Calibration and charged particle identification 129 E. Typical experiments on cluster decay modes 131 F. Other applications 132 References 134

122

I. Introduction Cluster radioactivity was predicted in 1980 by Sandulescu, P o e n a r u and Greiner [1], and the first model allowing the half-lives to be calculated was introduced a n d successively improved by P o e n a r u , Greiner et al. [ 2 - 8 ] T h e experimental discovery came four years later, when Rose a n d Jones [ 9 ] published their first identification of 1 4 C nuclei s p o n t a n e o u s l y emitted by 2 2 3 R a . T h e technique they used of the classical [10,11] Α Ε χ Ε telescope of semiconductor detectors [12,13], directly viewing the source, has been rapidly [ 1 4 , 1 5 ] o v e r t a k e n by two other m o r e selective methods, which will be presented in o u r chapter. An i m p o r t a n t d r a w b a c k of the m e t h o d used in [ref. 9 ] is the limited lifetime of Si detectors due to radiation d a m a g e . After irradiation with a b o u t 10 9 α / c m 2 the detector should generally be replaced. As shown in [ref. 16] the majority of trans-zirconium (Z > 40) nuclei, including even some "stable" ones, are metastable with respect to several cluster decay modes. Nevertheless, only the most favourable cases (half-lives shorter t h a n a b o u t 1026 s a n d branching ratios larger t h a n 1 0 " 1 6 ) are experimentally accessible with the presently existing detection techniques. These are generally observed for c o m b i n a t i o n s leading to a double magic d a u g h t e r nucleus (e.g. 2 0 8 P b or 1 0 0 S n ) or very close to it. U p to now, the

118

Ε. Hourany, I. Η. Plonski and D. N. Poenaru

e x p e r i m e n t a l l y c o n f i r m e d p a r e n t nuclei (see t h e review p a p e r s [ 1 7 - 2 2 ] a n d t h e references t h e r e i n ) e x h i b i t i n g cluster r a d i o a c t i v i t y a r e m a i n l y c o n f i n e d t o the t r a n s f r a n c i u m r e g i o n ( Z = 8 7 - 9 6 ) . T h e f o l l o w i n g m o d e s h a v e b e e n identified: 1 4 C d e c a y of 221 F r , 2 2 1 " 2 2 4 - 2 2 6 R a a n d 2 2 5 A c ; 2 0 O d e c a y of 2 2 8 T h ; 2 3 F d e c a y of 2 3 1 P a ; 2 4 N e d e c a y o f 230 32

T h )

23 1 p

a

a n d

238

232 - 2 3 4 34

U ;

2 8

M g

d e c a y

^ 2 3 ^

a n (

j 236,238pu; 3 0

M g

d e c a y

Q f

238p

u ;

242

S i d e c a y of P u , a n d S i r a d i o a c t i v i t y of C m . Half-lives in t h e r a n g e Ι Ο ^ - Ι Ο 2 6 s a n d b r a n c h i n g r a t i o s relative t o α - d e c a y of 1 0 " 1 6 - 1 0 ~ 9 h a v e b e e n m e a s u r e d . M o r e recently 1 2 C r a d i o a c t i v i t y of t h e p r o t o n rich n u c l e u s 1 1 4 B a , f o r w h i c h t h e d a u g h t e r is 102 S n , h a s b e e n o b s e r v e d (by T r e t y a k o v a et al. in J I N R D u b n a a n d b y G u g l i e l m e t t i et al. in a D a r m s t a d t - M i l a n o - B e r k e l e y c o o r p e r a t i o n ) in c o m p e t i t i o n w i t h ß-decay. T h i s is o n e r e p r e s e n t a t i v e of a n e w island of cluster e m i t t e r s p r e d i c t e d b y P o e n a r u , G r e i n e r a n d c o w o r k e r s [ 2 3 - 2 5 , 8 ] since 1984. It is i n t e r e s t i n g t o p o i n t o u t [ 18] t h a t in fact c l u s t e r r a d i o a c t i v i t y of 2 3 0 T h , 2 3 1 P a a n d U h a d a l r e a d y b e e n m e a s u r e d in the 1940s a n d 1950s, b u t w a s misidentified a s s p o n t a n e o u s fission. In f a c t t h e s p o n t a n e o u s fission half-lives of t h e s e nuclei are l o n g e r t h a n t h o s e t a b u l a t e d , w h i c h c o r r e s p o n d t o t h e p a r t i a l half-lives a g a i n s t 2 4 Ne decay. 232

T h e m a i n difficulty in this k i n d of e x p e r i m e n t is t h a t t h e useful e v e n t s a r e r a t h e r r a r e a n d m o r e o v e r t h e r e is a h u g e b a c k g r o u n d of a l p h a p a r t i c l e s ( b r a n c h i n g r a t i o s relative t o α - d e c a y l o w e r t h a n 10 ~ 9 ). H o w c a n o n e get rid of t h e s e a l p h a s ? T h e s o l u t i o n a d o p t e d in t h e e l e g a n t e x p e r i m e n t s p e r f o r m e d a t O r s a y , by H o u r a n i a n d his c o l l e a g u e s [ 1 4 , 2 6 , 2 7 , 1 7 , 2 0 ] , w a s t o deflect t h e u n w a n t e d 4 H e ions, simply o r d o u b l y ionized, b y a s t r o n g m a g n e t i c field p r o d u c e d in t h e s p e c t r o m e t e r S O L E N O [ 2 8 ] a n d t o select o n l y 1 4 C clusters t o r e a c h t h e d e t e c t o r in t h e f o c a l plane. W i t h this u n i q u e i n s t r u m e n t it w a s p o s s i b l e t o d i s c o v e r t h e fine s t r u c t u r e in cluster d e c a y [ 2 9 - 3 1 , 2 1 ] d i s c u s s e d b y M a r t i n G r e i n e r a n d W e r n e r Scheid [ 3 2 ] . An E n g e split-pole m a g n e t i c s p e c t r o m e t e r , w i t h a gas filled d e t e c t o r in its focal p l a n e , h a s b e e n used at A r g o n n e [ 3 3 , 3 4 ] t o c o n f i r m t h e m a s s n u m b e r of t h e e m i t t e d 1 4 C f r a g m e n t from 223Ra. A s e c o n d m e t h o d extensively used [ 1 5 , 3 5 , 1 8 , 1 9 , 3 6 - 3 9 , 2 2 ] a l l o w i n g b r a n c h i n g r a t i o s as low as 10 ~ 1 7 t o be m e a s u r e d , is b a s e d o n t h e solid s t a t e n u c l e a r t r a c k d e t e c t o r s ( S S N T D ) , [ 4 0 , 4 1 ] w h i c h a r e n o t sensitive t o a l p h a s a n d o t h e r l o w - Z particles, b e c a u s e they n e e d a c e r t a i n t h r e s h o l d of i o n i z a t i o n . T h e y a r e n o t expensive, b u t like p h o t o g r a p h i c plates, d o n o t deliver the i n f o r m a t i o n on-line; o n l y a f t e r a s u i t a b l e p o s t i r r a d i a t i o n c h e m i c a l e t c h i n g t h e t r a c k s are f o r m e d well e n o u g h . T h e e t c h i n g rate a l o n g the p a t h s of t h e i o n s d e p e n d s o n the c h a r g e n u m b e r of t h e i o n i z i n g particle. T h e p l o t of the e t c h i n g r a t e v e r s u s t h e residual r a n g e yields a Ζ i d e n t i f i c a t i o n . S S N T D h a v e b e e n , a n d c o n t i n u e t o be, widely used in a l a r g e v a r i e t y of n u c l e a r p h y s i c s e x p e r i m e n t s , [ 4 1 , 3 7 , 4 2 ] p a r t i c u l a r l y f o r r a r e e v e n t s e n c o u n t e r e d in s p o n t a n e o u s fission [ 4 3 , 4 4 ] , cold fission, [ 4 5 ] a n d s p o n t a n e o u s l y fissioning s h a p e i s o m e r s [ 4 6 - 4 9 ] . I n t h e next s e c t i o n w e shall p r e s e n t t h e c o n s t r u c t i o n a n d t h e m a i n c h a r a c t e r i s t i c s of S O L E N O , t h e p r e p a r a t i o n of r a d i o a c t i v e s o u r c e s , t h e d e t e c t i o n s e t u p , a n d s o m e typical e x p e r i m e n t s o n cluster r a d i o a c t i v i t y a n d its fine s t r u c t u r e . T h e n we shall c o n t i n u e by d e s c r i b i n g t h e m o s t i m p o r t a n t p h y s i c a l processes t a k i n g place in a S S N T D a n d finally l e a d i n g t o a t r a c k w h i c h c a n b e seen w i t h a m i c r o s c o p e , h o w o n e m a y d e r i v e t h e c h a r a c t e r i s t i c s of t h e i n c o m i n g p a r t i c l e f r o m t h e s h a p e a n d d i m e n s i o n s of t h i s t r a c k , h o w t o c h o o s e a s u i t a b l e t y p e of d e t e c t o r t o fulfil t h e

Track Detectors and Soleno Applied for Cluster Reactivities

119

requirements of an experiment a n d which performances can be expected, what kind of radiation d a m a g e prevents a long-time exposure, etc. In spite of the fact that we illustrate our presentation with specific applications for cluster decay measurements, the extentions to other kinds of rare events can easily be made.

II. The magnetic spectrometer SOLENO T h e separation of 1 4 C ions a n d the high resolution m e a s u r e m e n t of their kinetic energy spectrum (essential in experiments on the fine structure) has been achieved with the S O L E N O spectrometer at I P N Orsay, F r a n c e [28], T h e main q u a n t i t y of interest for selection is the magnetic rigidity Bp. T h e magnetic field focuses 1 4 C 6 + ions while rejecting the α + and a + + particles emitted by the same source. An Ε or Δ £ χ Ε detector is m o u n t e d in the focal plane. N o t only is the detector protected against the d a m a g e by α-particles, but also the r u n n i n g time can be considerably shortened (e.g. five days c o m p a r e d to six m o n t h s (for a detection system placed in direct view of the source)) by using a much stronger source.

A. Superconducting solenoidal spectrometer T h e magnetic field of S O L E N O is obtained with a solenoidal s u p e r c o n d u c t i n g coil placed in a cryostat s u r r o u n d e d by an iron shield (see Fig. 1). T h e source a n d the

1

Fig. 1. The general setup of SOLENO used in 1 4 C radioactivity measurements. The system has a cylindrical symmetry. 1-iron shield, 2-solenoidal coil, 3-vacuum chamber, 4-source, 5-iris, 6-obturators, 7-detector, 8-turbo pump, 9-cryogenic pump, 10-valve. (Reprinted with permission from Hourani, E., Rosier, L., Berrier-Ronsin, G., Elayi, Α., Mueller, A. C., Rappenecker, G., Rotbard, G„ Renou, G., Liebe, Α., Stab, L„ and Ravn, H. L„ Phys. Rev. C, 44,1424, © 1991, The American Physical Society).

120

Ε. Hourany, I. Η. Plonski and D. N. Poenaru

detector system are mounted on the axis of the internal coaxial cylindrical vacuum chamber possessing a perfect azimuthal symmetry. Compared to the above mentioned Enge splitpole spectrometer, the solid angle of S O L E N O (defined by the iris) is much larger: 200 msr instead of 5 msr. The shutter prevents α-particles, emitted by the source along the symmetry axis (which are not deflected by the axial magnetic field), from reaching the detector. The experiment is performed in the following manner [31,21], First, the strong source (to investigate the 1 4 C emission) and a weak (1/1000) one of equal diameter and from the same material (to calibrate the S O L E N O ) are measured. Each source is placed under vacuum in direct view of a silicon detector with a known aperture, at a sufficiently large distance (say 2 m) so as the α-particle counting rate is tolerable. Then after filling the cryostat with liquid helium, the electric current intensity, /, through the coil is set to focus the 1 4 C 6 + ions on the silicon semiconductor detector. F o r this setting the doubly ionized α + + particles are focused well before the detector, while the singlecharged α + are deflected. The focusing efficiency for a given geometry of source and detector is characterized with a help of the transmission curve (Fig. 2), which is a plot of the effective solid angle Ω versus the single parameter y = Bp/1, where Bp is the magnetic rigidity of the focused ion and / is the current of S O L E N O . The transmission curve is established by counting the α + + particles of a chosen α-line (fixed Bp) from the weak source, for several values of the current /. O n e can choose thereafter the location y0 of 1 4 C ion with a given rigidity (Bp) 0 by setting the current, / 0 , such that (Bp)0/I0 =y0. In 1 4 C radioactivity the parent and emitted nuclei are in their ground state. D u e to the energy balance of the decay, a measured value of the kinetic energy of 1 4 C ions corresponds to a given state (ground state or excited) of the daughter nucleus and determines its location on the transmission curve. In Fig. 2, used in experiments on the fine structure [31], the arrows give the locations of the ground state and a few low-lying excited states of the residual nucleus, 2 0 9 P b and of 2 1 0 P b , for several values of the electric current intensity in the coil of S O L E N O .

B. Sources The sources used in cluster decay measurements usually deliver a number of unwanted α-particles several ( > 9) orders of magnitude larger. The limit in source intensity is dictated by its thickness which should not degrade the kinetic energies of emitted particles too much and should not exaggeratedly increase the straggling width in energy. 14 C emitters are mainly Ra isotopes, because the maximum emission rate occurs when the daughter is the doubly magic 2 0 8 P b nucleus or a closely neighbouring one. The Ra isotopes are available as members of the natural decay chains of Uranium and Thorium or can be produced in nuclear reactions by fragmentation of Uranium or Thorium targets. Intense sources are needed to compensate for the low branching ratio 1 4 C / a of (10" l o - 1 0 " 1 2 ) . Two main types of sources have been used so far: 1) chemically separated and deposited sources [30], and 2) ion-implanted sources [21].

Track Detectors and Soleno Applied for Cluster Reactivities

Bp/I

121

(Txm/A)

Fig. 2. A typical transmission curve of S O L E N O in fine structure measurements, for a single Si detector of 450 m m 2 in the focal plane. The effective entrance solid angle is plotted versus the ratio Βρ/1, where Bp is the magnetic rigidity of the ions emitted by the source and / is the electric current intensity through the coil. T h e locations of the g r o u n d state and a few low-lying excited states of the residual nucleus in the decay of 2 2 3 R a and 2 2 4 R a are labeled with numbers and letters respectively. (Reprinted with permission from H o u r a n i , E., Rosier, L., Berrier-Ronsin, G., Elayi, Α., Mueller, A. C., Rappenecker, G., Rotbard, G., Renou, G., Liebe, Α., Stab, L., and Ravn, H. L„ Phys. Rev. C, 44,1424, © 1991, The American Physical Society).

Examples from the first category are two members of the decay chain of 2 3 1 P a : 2 2 7 Ac (T = 21.77 y) and 2 2 7 Th(T = 18.72 d), which generate 2 2 3 R a ( 7 = 11.43 d). An activity of 210 μ 0 (about 300 times stronger than that used by Rose and Jones) of 2 2 3 Ra has been obtained [14] with a thin deposit of 2 2 7 Ac having a diameter of 8 mm. A radioactive deposit of 2 3 0 U (T = 20.8 d) was used to generate an activity of 600 μ(Ιί of 2 2 2 R a ( T = 38 s). The 2 3 0 U itself was obtained by means of a chemical separation from the products of the reaction 2 3 2 Th (p, 3n) with Ep = 34 MeV from the cyclotron of Orleans. The thicker (1.8 mg/cm 2 ) source [26] of 2.5 mCi of 2 2 6 Ra ( T = 1600y) has been deposited within a S 0 4 R a or CIRa compound. From all three kinds of sources mentioned so far an isotope of Rn is produced as it is a member of the respective decay chain. This gas is emanated from the source, travels inside the vacuum chamber and activates the environment of the detector with its daughters, eventually producing background counts. High quality implanted sources of Ra isotopes have been made [31] by means of ISOLDE mass separator at CERN, Geneva. A 2.8 μΑ proton beam of 600 MeV accelerated in the synchrocyclotron bombarded a 55 g/cm 2 thick thorium carbide

122

Ε. Hourany, I. Η. Plonski and D. N. Poenaru

target. A typical p r o d u c t i o n rate of 8 10 9 ions/sec has been obtained for 2 2 4 R a . The ions of 60 keV extracted f r o m the separator were implanted into a catcher. Recently, a source of 2 2 3 R a was produced [21] with the new I S O L D E operating at p r o t o n s y n c h r o t r o n of 1 G e V of C E R N . A production rate similar to the one previously achieved at the synchrocyclotron was obtained. By using the spetrometer S O L E N O with a good detector, the quality of the results is mainly governed by the quality of the radioactive source. O n the one hand, the sources produced at I S O L D E possess thin windows. Actually, the implantation goes within a depth of 2 0 0 - 4 0 0 Ä. T o decrease the sputtering in the implantation, the b e a m spot is regularly moved over an area of φ χ 8 m m of the c a r b o n catcher. In order to overcome the emission of recoil Rn a t o m s and to allow the 1 4 C ions to reach the charge state equilibrium along their p a t h s in the layer, it was estimated that it is convenient to deposit an Al layer of 1000-2000 Ä. Indeed, with these precautions implanted sources of spectroscopic quality have been made. T h e first experiment with 224 R a gave [31] an energy resolution of 150 keV with no energy tailing a n d the second experiment with 2 2 3 Ra gave [21 ] a perfect result of 90 ke V of energy resolution with no tailing a n d n o b a c k g o r u n d . O n the other hand, in spite of their historical i m p o r t a n c e for the discovery of the 1 4 C radioactivity and of the fine structure, the chemically separated sources [30] present a d r a m a t i c d r a w b a c k . W h e n they are strong, they contain m u c h material causing finite thickness a n d noticeable inhomogeneity. They fail in measuring precise spectroscopic factors, not only because the energy resolution is p o o r b u t also because they show: (i) a low energy tailing m a k i n g it difficult to measure small size peaks falling on the tail of large size ones, (ii) spurious peaks appearing on the low energy side of p r o n o u c e d peaks.

C. Detection system Either £ or a AE χ £ telescope of silicon surface barrier detectors m a d e at Orsay have been placed in the focal plane of S O L E N O . T h e thin transmission detector Δ £ is a b o u t 9 μηι thick with an area of 200 m m 2 . T h e total energy £ detector has a larger area (300 m m 2 ) and thickness (200 μηι). With a s t a n d a r d electronic setup [14], three p a r a m e t e r s (Ε,ΑΕ,Τ) have been recorded on a magnetic t a p e event by event. Τ is the time delay between the two other signals given as an o u t p u t of a time-to-amplitude converter. T h e amplitude thresholds in the two energy channels have been adjusted to limit the counting rate of the stored events to a few counts per second. In a A £ versus £ plot (Fig. 3), the α and 1 4 C particles fall o n t o different hyperbolalike locations, a n d in such a way they can be identified [31].

D. Experiments of cluster radioactivity and its fine structure T h e O r s a y scientists [14] gave with S O L E N O the first confirmation of the experiment performed by Rose and Jones on 1 4 C radioactivity of 2 2 3 R a . Also the same kind of decay of 2 2 2 R a and 2 2 6 R a has been investigated [26].

Track Detectors and Soleno Applied for Cluster Reactivities

U

C

from

223

Ra{

123

1 event • 10 events

C or , 6 0 b e a m f - 1 / 5 M a x . contour l i m i t s 1 - 1/50Max. ,6 0 35MeV ,2

,6 LU 50 _ 0 j 2 5 M e V




Ne

CD

--

4

>

F 3

2

0

2

4

6

8

1 0

12

14

16

18

20

22

Range (μηι) Fig. 6. Identification of Ne and Mg clusters spontaneously emitted from 2 3 4 U (Reproduced with permission, from Price, P.B. Annu. Rev. Nucl. Part. Sei., 39,19, © 1989, by Annual Reviews Inc.).

E. Typical experiments on cluster decay modes The evidence for l 4 C emission f r o m 222 Ra and 2 2 4 R a together with a n o t h e r c o n f i r m a t i o n from 2 2 3 R a was given [15] using ion-implanted sources at I S O L D E a n d S S N T D . T h e Ra ions were implanted at the b o t t o m of a c u p whose t o p a n d walls were covered by polycarbonate foils consisting of a 125 μηι Tuffak sheet covered w i t h a 10 μηι M a k r o f o l sheet; in a subsequent experiment R o d y n e - P p o l y c a r b o n a t e was used. Tens of events were counted. T h e results are presented as range histograms a n d the Ζ identification is achieved in a bidimensional plot of the etching rate versus the residual range. In Fig. 7 one can see a histogram of the range distributions measured [55] for 2 4 N e nuclei spontaneously emitted from 2 3 2 U . T h e ranges calculated f r o m the ß - v a l u e s for emission of different isotopes of N e are indicated with arrows. T h e C r o n a r plates were exposed for a m o n t h in a hemispherical geometry a r o u n d a 0.5 m C i source of 2 3 2 U . Only a m o d e r a t e v a c u u m is necessary. Usually the polyester films present optical nonuniformities, hence the technique of silicon replicas has been used. Both the diameters a n d the lengths of the tracks have been measured on these replicas of the detector surface m a d e after etching. U n f o r t u n a t e l y due to a source c o n t a m i n a t i o n with 252 C f , the s p o n t a n e o u s fission half-life of 2 3 2 U could not be measured in 1984. In 1990 a new m e a s u r e m e n t was reported by the M i l a n o g r o u p w o r k i n g with a p h o s p h a t e glass detector [38]. They gave an u p p e r limit for the s p o n t a n e o u s fission half-life of a b o u t 100 times longer t h a n that obtained in 1950, proving that in fact the N e decay had been measured in 1950 and for 40 years this was misinterpreted as due to spontaneous fission.

132

Ε. Hourany, I. Η. Plonski and D. N. Poenaru 15

—V 20

•

| 10 φ ο

232

ω

u

23 22 24

V

I

Ne + Pb

0

0

20

30

40

Range (/vm) Fig. 7. The histogram of the range distributions measured for 2 4 N e nuclei spontaneously emitted from 2 3 2 U . The ranges calculated from the Q-values for emission of different isotopes of Ne are indicated with arrows (Reprinted with permission from Barwick, S. W., Price, P. B., and Stevenson, J. D. Physical Review, C 31, 1984, © 1985, The American Physical Society).

The first nuclide for which three groups of fission-like decay-modes (cold fission with Zr the most probable light fragment, α-decay, and three cluster radioactivities 24,26 N e and 2 8 M g ) have been experimentally detected is 2 3 4 U . In this way one has a beautiful confirmation of the validity of a unified treatment of these processes in a wide range of mass asymmetry [45]. Phosphate glass detectors sensitive to clusters with Ζ > 8 in a planar geometry have been used [35]. In Fig. 6 we present the measured 108 Ne and 36 M g decays in the Berkeley-Livermore experiment performed in 1989. The lowest branching ratio ever measured is b = 6 - 1 0 " 1 7 , and it was obtained [56] for 2 8 3 0 M g decay of 2 3 8 P u by using the phosphate glass LG-750 in a long exposure to a 10.5 mg source. Fission fragments were stopped in a 13 μιη Al foil covering the detectors. loo

F. Other applications Besides the above mentioned experiments on cluster decay modes, there are many other fields of nuclear physics, astrophysics, and geophysics, where the S S N T D have been successfully applied [41, 42, 37] For α-decay they are of limited use, because the majority of S S N T D are not sensitive to He isotopes. Even when they are, as for instance, polycarbonate and cellulose nitrate, it is difficult to recognize the tracks because vT is not very different from vG and the range of accepted energies is very small: 0.2 to 3 MeV in the former case and 0.1 to 4 - 6 MeV in the latter. Sometimes S S N T D are used to determine with good resolution (of the order of 1 μπι) the spatial distribution of an α emitter contained in a target. When such an α sensitive detector is deposited in air it can accumulate tracks from the decay of radon.

Track Detectors and Soleno Applied for Cluster Reactivities

133

The detector materials with low concentration of uranium impurity (polymers, quartz or phosphate glasses) are frequently used to measure spontaneous fission half-lives or the activity and distribution of fissioning nuclei in some targets. They have been also employed to measure the kinetic energy and the mass distribution of fission fragments. M a n y experiments on spontaneously fissioning shape isomers [46,57] have been performed with track detectors. By assuming that the predicted superheavy elements would exhibit spontaneous fission, S S N T D have been extensively used [37] to search for such nuclei in nature or in nuclear reactions attempting to synthesize them. O n e idea for the search in nature consisted in the investigation of some very old crystals and glasses sensitive to fission fragments, in which the fission of a superheavy nucleus could leave some number of tracks accumulating in time. Only very low upper limits have been observed in such experiments and in similar ones looking for a ternary or α-accompanied fission. The heaviest (transfermium) nuclei, produced at J I N R D u b n a by heavy-ion fusion reactions with very low cross-sections, have been identified by measuring their spontaneous fission half-lives determined with track detectors. Nuclear reaction mechanisms and the identification of charged particles in their exit channels have been also benefitted by the use of SSNTD. Low induced fission cross-sections, angular distributions of fission fragments, determination of the mean number of neutrons emitted per fission event, are few examples of such studies. Detectors based on poly-allyl-glycol-carbonate have been used for α-decaying nuclei. Some polymeric detectors lose their etchability when operated in vacuum, due to the absence of oxygen. In this case they are restored by irradiation with ultraviolet ( λ < 3 1 0 η π ι ) light. The muscovite mica detectors give good results even in a high temperature (400°C) and chemically agressive (volatile chlorides) environment. Very short half-lives (in the range 1 0 " 1 8 to 1 0 " 1 5 s) of some nuclear states have been determined by using the "shadow effect" (see the chapter by Lieb in this book) observed when single-crystal targets (e.g. 1 8 6 W, 1 8 1 T a or 2 3 8 U ) are bombarded with heavy ion beams, also with the help of SSNTD. The list of many other applications of S S N T D in nuclear physics experiments is very large. One can mention: some reactions of astrophysical interest induced on light nuclei; the elastic scattering of charged particles from low intensity radioactive secondary beams (see the chapter by Tanihata in the present book); the multifragmentation of the projectile in relativistic heavy-ion collisions (see the chapter by Coffin in this book), etc. Among the applications, the production of special filter membranes with a precisely determined pore diameter (in the range 0.01 to 10 μηι), able to withstand chemical solutions with pH from 2 to 12, used in biotechnology, microbiology, pharmaceutic, milk, and wine industries or in the ultraclean workshops where microelectronic devices are made, is of particular interest. In this case a plastic foil (e.g. a 10 μηι thick PTE) is uniformly irradiated with heavy ions of given energy. The position of the beam spot on the foil surface is changed in a similar way to that of an electron beam in a TV cathode-ray tube, by an electromagnetic system of deflection, and the foil is slowly moved with a constant velocity in order to get the desired density of the pores after etching.

134

Ε. Hourany, I. Η. Plonski and D. N. Poenaru

Acknowledgements This work has been supported by the Institut de Physique Nucleaire, Orsay, Bundesministerium für Forschung und Technologie, the Deutsche Forschungsgemeinschaft, Bonn, and the Institute of Atomic Physics, Bucharest. One of us (DNP) has received a bourse haut niveau from the Ministere de l'Enseignement Superieur et de la Recherche, Paris and a donation of computer and copying equipment from the Soros Foundation for an Open Society.

References [1] Sändulescu, Α., Poenaru, D. N. and Greiner, W., Soviet Journal Particles and Nuclei 11, 528-541 (1980). [2] Poenaru, D. N. and Iva§cu, M. Report NP-17, Central Institute of Physics, Bucharest (1980). [3] Poenaru, D. N., Iva$cu, M., Sändulescu, Α., and Greiner, W., Journal of Physics G: Nuclear Physics 10, L183-L189 (1984). [4] Poenaru, D. N., Iva$cu, M., Sändulescu, Α., Greiner, W., Physical Review, C32, 572-581 (1985). [5] Poenaru, D. N., Greiner, W., Iva§cu, M., Mazilu, D., and Plonski, I. H. Zeitschrift für Physik, A 325, 435-439 (1986). [6] Poenaru, D. N., Iva§cu, M., Mazilu, D., Iva$cu, I., Hourani, E., and Greiner, W. In Proc. International Symposium on Developments in Nuclear Cluster Dynamics, Sapporo, K. Akaishi, K. Kato, H. Noto, and S. Okabe, editors, 76-87, (World Scientific, Singapore, 1989). Poenaru, D. N., Schnabel, D., Greiner, W., Mazilu, D. and Gherghescu, R. Atomic Data and Nuclear Data Tables 48, 231-327 (1991). [7] Poenaru, D. N., Hourani, E., and Greiner, W. Annalen der Physik 3, 107-117 (1994). [8] Poenaru, D. N. and Greiner, W. In Handbook of Nuclear Properties, Poenaru, D. N. and Greiner, W., editors, chapter 5, 131-182, Oxford University Press (1996). [9] Rose, H. J. and Jones, G. A. Nature 307, 245-247 (1984). [10] Goulding, F. S. and Harvey, B. G. Annual Review of Nuclear Science 25, 167-240 (1975). [11] Pougheon, F. Chapter in the present book. [12] Poenaru, D. N. and Vilcov, N. Measurement of Nuclear Radiation with Semiconductor Detectors. Chemical Publishing Company, New York (1969). [13] Ponpon, J. P. Chapter in the present book. [14] Gales, S., Hourani, E., Hussonnois, M., Schapira, J. P., Stab, L. and Vergnes, M. Physical Review Letters 53, 759-762 (1984). [15] Price, P. B., Stevenson, J. D., Barwick, S. W. and Ravn, H. L. Physical Review Letters 54, 297-299(1985). [16] Poenaru, D. N., Greiner, W., Iva$cu, M., and Sändulescu, A. Physical Review, C 32, 2198-2200(1985). [17] Hourani, E. and Hussonnois, M. In Particle Emission from Nuclei, Vol. II: Alpha, Proton and Heavy Ion Radioactivities, Poenaru, D. N. and Iva^cu, M., editors, chapter 6,171-187. CRC Press, Boca Raton, Florida (1989). [18] Price, P. B. and Barwick, S. W. In Ref. 17, Vol. II, chapter 8, pp. 205-231. [19] Ha§egan, D. and Tretyakova, S. P. In Ref. 17, Vol. II, chapter 9, pp. 234-257. [20] Hourani, E., Hussonnois, M., and Poenaru, D. N. Annalesde Physique (Paris) 14,311-345 (1989). [21] Hourany, E. In Nuclear Decay Modes, edited by D. N. Poenaru, Fundamental and Applied Nuclear Physics Series, lOP, Bristol and Philadelphia, pages 350 (1996). [22] Bonetti, R. and Guglielmetti, A. In Nuclear Decay Modes, Ref 21, pages 370.

Track Detectors and Soleno Applied for Cluster Reactivities

135

[23] Poenaru, D. N. and Iva§cu, M. In Proc. International Summer School on Atomic and Nuclear Heavy-Ion Interactions, Poiana Brasov, 277-331 Central Institute of Physics, Bucharest (1984). [24] Poenaru, D. N., Iva§cu, M., Sändulescu, Α., and Greiner, W. Report NP-46, Central Institute of Physics, Bucharest (1985). Greiner, W., Iva$cu, M., Poenaru, D. N., and Sändulescu, A. In Τ reatise on Heavy Ion Science, Vol. 8, Bromley, D. A. editor, 641-722, Plenum Press, New York, (1989). [25] Poenaru, D. N., Greiner, W., and Gherghescu, R. Physical Review, C 47,2030-2037 (1993). [26] Hourani, E., Hussonnois, M., Stab, L., Brillard, L., Gales, S., and Schapira, J. P. Physics Letters, Β 160, 375-379 (1985). [27] Hourani, E. In Lecture Notes in Physics, 279,383-398. Springer Verlag, Heidelberg, (1987). [28] Schapira, J. P., Azaiez, F., Fortier, S., Gales, S., Hourani, E., Kumpulainen, J., and Maison, J. M. Nuclear Instruments and Methods, 224, 337 (1984). [29] Brillard, L., Elayi, A. G., Hourani, E., Hussonnois, M., Le Du, J. F., Rosier, L. H. and Stab, L. C. R. Acad. Sei. Paris 309, 1105-1110 (1989). [30] Hussonnois, M., Le Du, J. F., Brillard, L., Dalmasso, J., and Ardisson, G. Physical Review, C 43, 2599-2604(1991). [31] Hourani, E., Rosier, L., Berrier-Ronsin, G., Elayi, Α., Mueller, A.C., Rappenecker, G., Rotbrad, G., Renou, G., Liebe, Α., Stab, L., and Ravn, H. L. Physical Review C44, 1424-1434 (1991) and Physical Review C52, 267-270 (1995). [32] Greiner, M. and Scheid, W. Journal of Physics G: Nuclear Physics 12, L229-L234 (1986). [33] Kutschera, W., Ahmad, I., III, S. G. Α., and J. E. Gindler, A. M. F., Henning, W„ Ishii, T„ Paul, M„ and Rehm, Κ. E. Physical Review, C 32, 2036-2042 (1985). [34] Henning, W. and Kutschera, W. In Ref. 17, Vol. II, chapter 7, pp. 189 204. [35] Price, P. B. Annual Review of Nuclear and Particle Science 39, 19-42 (1989). [36] Tretyakova, S. P., Zamyatnin, Y. S., Kovantsev, V. N., Korotkin, Y. S., Mikheev, V. L., and Timofeev, G. A. Zeitschrift für Physik, A 333, 349-353 (1989). [37] Tretyakova, S. P. Soviet Journal of Particles and Nuclei 23, 156-186 (1992). [38] Bonetti, R., Fioretto, E., Migliorino, C., Pasinetti, Α., Barranco, F., Vigezzi, E., and Broglia, R. A. Physics Letters Β 241, 179-241 (1990). [39] Bonetti, R., Chiesa, C., Guglielmetti, Α., Migliorino, C., Cesana, Α., Terrani, M. and Price, P. B. Physical Review, C 44, 888-890 (1991). [40] Price, P. B. and Walker, R. M. Journal of Applied Physics 33, 3400 (1962). [41] Fleischer, R. L., Price, P. B., and Walker, R. M. Nuclear Tracks in Solids. Principles and Applications. University of California Press, Berkeley (1975). [42] Durrani, A. and Bull, R. K. Solid State Nuclear Track Detectors. Principles Methods and Applications. Pergamon Press, Oxford (1987). [43] Hoffman, D. C. and Somerville, L. P. In Ref. 17, Vol. Ill: Fission and Beta-Delayed Decay Modes, chapter 1, pp. 1-40. [44] Hoffman, D. C., Hamilton, Τ. M., and Lane, M. R. In Nuclear Decay Modes, Ref 21, page 393. [45] Poenaru, D. N., Iva$cu, M., and Greiner, W. In Ref. 17, Vol. Ill, chapter 7, pp. 203-235. [46] Poenaru, D. N. Annales de Physique (Paris) 2,133-168 (1977). [47] Metag, V., Habs, D., and Specht, Η. J. Physics Reports 65, 1 (1980). [48] Bj0rnholm, S. and Lynn, Κ. E. Review of Modern Physics 52, 725 (1980). [49] Poenaru, D. N., Iva§cu, M., and Mazilu, D. In Ref. 17, Vol. Ill, chapter 2, pp. 41-61. [50] Sheline, R. K. and Ragnarsson, I. Physical Review, C 43, 1476-1479 (1991). [51] Benton, Ε. V. and Nix, W. D. Nuclear Instruments and Methods 67, 343 (1969). [52] Price, P. B., Cook, L. M., and Marker, A. Nature (London) 325, 137 (1987). [53] Henke, R. P. and Benton, Ε. V. Nuclear Instruments and Methods 97, 483 (1971). [54] Somogyi, G. Nuclear Instruments and Methods 173, 21 (1980). [55] Barwick, S. W., Price, P. B., and Stevenson, J. D. Physical Review, C 31,1984-1986 (1985). [56] Wang, S., Snowden-Iffi, D., Price, P. B., Moody, K. J., and Hulet, Ε. K. Physical Review, C39, 1647-1650(1989). [57] Poenaru, D. N. and Plonski, I. H. In Nuclear Decay Modes, Ref 21, page 433.

5

New Generation of Gamma-Detector Arrays

Rainer Μ.

Lieder

Table of Contents I. Introduction 137 A. Physics motivation 137 B. Population of highly-excited nuclei 141 1. Nuclear reactions for populating states of high spin and high excitation energy 141 2. Deexcitation of highly-excited compound nuclei 143 C. In-beam γ-spectroscopy techniques 143 II. Performance of gamma-detector arrays 148 A. Figure-of-merit 149 B. Energy resolution 153 C. Total photopeak efficiency and peak-to-total ratio 155 D. Features of y-detector arrays 157 III. High-resolution gamma-ray spectrometers 160 A. Escape-suppressed germanium detectors 161 B. Composite detectors 164 1. The CLOVER detector 165 2. The CLUSTER detector 167 3. The use of CLOVER and CLUSTER detectors as Compton Polarimeters 172 C. Segmented detectors 173 IV. Gamma-detector arrays 175 A. Present y-detector arrays 177 1. The GASP array 177 2. The EUROGAM II array 178 3. The GAMMASPHERE I array 181 B. Future γ-detector arrays 181 1. The full GAMMASPHERE array 182 2. The EUROBALL III array 183 C. Comparison of γ-detector arrays 185 References 185

I. Introduction A. Physics motivation T h e study of high-spin states in nuclei was initiated in 1963 by M o r i n a g a a n d G u g e l o t [1] by the observation that discrete γ-rays can be detected in the deexcitation of nuclei produced by nuclear reactions. They carried out the first " i n - b e a m " γ-spectroscopic experiments utilizing Nal(Tl) detectors to m e a s u r e the γ-radiation emitted p r o m p t l y

138

Rainer Μ. Lieder

f r o m the target. In this pioneer experiment M o r i n a g a a n d G u g e l o t [ 1 ] could only identify γ - t r a n s i t i o n s with a n intensity of « 1 (normalized to the s u m of t h e intensities of all t r a n s i t i o n s feeding t h e g r o u n d state) d u e to t h e p o o r energy resolution of Nal(Tl) detectors, being 7 . 7 % in their experiments. T h e d e v e l o p m e n t of i n - b e a m γ-spectroscopy since then is related to t h e c o n t i n u o u s i m p r o v e m e n t of the features of γ-ray spectrometers, viz. the energy resolution, peak-tototal r a t i o a n d total p h o t o p e a k efficiency, as depicted in Fig. 1 for a γ - r a y energy of 1332 keV. T h e energy resolution was i m p r o v e d by a f a c t o r of > 15 by t h e i n t r o d u c t i o n of G e d e t e c t o r s [ 2 ] in 1967. T h e peak-to-total ratio was increased by a f a c t o r of > 5 by the i n t r o d u c t i o n of escape-suppressed G e detectors in 1980 into i n - b e a m γ-spectroscopy [ 3 , 4 ] , In G e d e t e c t o r s only a f r a c t i o n of the γ-rays are fully a b s o r b e d a n d c o n t r i b u t e to the p h o t o p e a k . T h e r e m a i n i n g γ-rays are scattered o u t of t h e G e detector mainly d u e t o the C o m p t o n effect, d e p o s i t i n g only a f r a c t i o n of their energy. T h e y p r o d u c e a n u n w a n t e d b a c k g r o u n d which can be reduced if the G e detector is s u r r o u n d e d by a n escape-suppression d e t e c t o r which detects the C o m p t o n - s c a t t e r e d γ-rays so t h a t they can be suppressed electronically by m e a n s of an anti-coincidence circuit. T h e escape-suppression detectors were initially large N a l ( T l ) scintillation

Fig. 1. Development of the features characteristic for γ-ray spectrometers at Ε = 1332 keV, viz. the energy resolution AEy, peak-to-total ratio P/T and total photopeak efficiency P ph since 1963, when the first in-beam γ-spectroscopy experiment was performed.

New Generation of Gamma-Detector Arrays

139

detectors and subsequently the more efficient, hence smaller, BGO (bismuth germanate) scintillators with holes in which the Ge detectors are inserted [3]. The introduction of BGO detectors enabled the solid angle occupied by an escape-suppressed Ge detector to be reduced considerably. The increase of the total photopeak efficiency (total detection efficiency for full-energy γ-rays) by a factor of > 50 represents the most recent progress [4,5,6] This was accomplished by the increase of the number of escapesuppressed Ge detectors in the array. The first generation of γ-detector arrays consisted of 6 - 2 1 escape-suppressed Ge detectors with total photopeak efficiencies of P p h ~ 0.5-1.5% for γ-rays of 1.33 MeV. Worldwide about a dozen arrays of this size have been built, e.g., TESSA3 (U.K.), Chateau de Cristal (France), OSIRIS (Germany), NORDBALL (Denmark), HERA (USA) and the 8π Spectrometer (Canada) [6,7], The present second generation of γdetector arrays consists of 30-54 escape-suppressed spectrometers with total photopeak efficiencies of Pph χ 3 - 8 % for γ-rays of 1.33 MeV. They are E U R O G A M I and II (U.K./France), GASP (Italy) and GAMMASPHERE I (USA) [6], The improvement of detector technology can be visualized by comparing the spectrum of the Yrast band in 1 5 6 Dy (The Yrast band is formed by the states of lowest excitation energy for a given spin) obtained in 1963 by Morinaga and Gugelot [1] using a Nal(Tl) detector with that of 1 5 8 Er obtained recently by Simpson et al. [8] with the modern γ-detector array EUROGAM II as shown in Fig. 2. In the former spectrum only four strong transitions can be seen whereas in the latter case the Yrast sequence could be identified up to a spin of / = 47. The progress in detector technology is reflected in important discoveries, as exemplified by two examples: In 1971 the band-crossing or backbending phenomenon [9] was found following the employment of two high-resolution Ge detectors as coincidence spectrometers. The band-crossing effect occurs in rotating deformed nuclei and results from the crossing and interaction of two rotational bands. The intersecting bands have configurations differing by one pair of particles which is broken up at the crossing and whose angular momenta are aligned along the rotation axis due to strong Coriolis forces in the fast rotating nucleus [10]. The rotation alignment of a particle pair is connected with a decrease of the γ-ray energies of the rotational transitions in the band-crossing region as can be seen in the lower portion of Fig. 2. The term backbending effect is related to this observation. In 1986 superdeformed bands in medium-heavy nuclei at high spins were discovered [11] after the introduction of arrays of escape-suppressed Ge detectors since their intensity is —I ΙΟ5

TT

1 Γ"r-

52 MeV

2+ l0 *

'

l56Gd(a,4n)l56Dy

C

•:··:•·••:..·••····•::•·.

β

„+ N,.

I03

10

20

WOO -

ι

ι

I ι ι ι ι I ι 50 100 Channel Number

lJ_i ι L l 150 200

'2s T e ( , Α-Π)'58 Er

vi«alignment

165 MeV '

TCh||/2 alignment-

3000 ιη c 3 2000

1000

500

1000

1500

EY (keV) Fig. 2. Comparison of spectra obtained by Morinaga and Gugelot, 1963 with one Nal(Tl) detector and by Simpson et al., 1994 with the modern γ-detector array EUROGAMII. The Yrast transitions in 1 5 6 Dy (upper portion) and in 158 Er (lower portion) are shown, respectively.

trometers. Two arrays are presently under construction, viz. The European γ-detector array E U R O B A L L III and the American γ-detector array G A M M A S P H E R E . They will allow nuclei to be studied under extreme conditions close to the limit of stability for which the excitation energy, the angular momentum or the ratio of protons to neutrons take extreme values. In this way nuclear models can be further tested. A few examples may be given:

New Generation of Gamma-Detector Arrays

141

• Search for nuclei of exotic shapes at very large angular m o m e n t a , e.g., of hyperdeformed shape with an axis ratio of « 3 : 1 . • Study of highly excited nuclei by means of giant resonances to search for fluctuations in shape a n d orientation a n d a possible liquid to gas phase transitions. • Investigation of the features of medium-heavy nuclei with the same p r o t o n a n d neutron n u m b e r u p to 1 0 0 S n . They lie close to the p r o t o n - d r i p line a n d can hence be populated only very weakly.

B. Population of highly-excited nuclei T h e design of γ-detector arrays for the study of nuclei in states of high spin a n d high excitation energy has to t a k e into consideration the way they are p o p u l a t e d . It is therefore necessary to discuss the p o p u l a t i o n a n d deexcitation m e c h a n i s m in detail.

1. Nuclear reactions for populating

states of high spin and high excitation

energy

Nuclear reactions are p r e d o m i n a n t l y used to p o p u l a t e nuclear states of high spin a n d high excitation energy. T h e type of reaction depends on the masses of projectile and target, the energy of the projectile a n d the impact parameter. F o r energies of 4 - 5 MeV/nucleon fusion reactions occur essentially for all impact p a r a m e t e r s smaller t h a n the distance of closest a p p r o a c h . F u s i o n reactions allow the highest possible angular m o m e n t u m I with the largest possible cross section σ to be transferred. In this way mainly neutron-deficient nuclei are p r o d u c e d as will be explained subsequently. F o r grazing collisions various types of reactions occur depending on the projectile energy and the mass asymmetry of the projectile-target system. If the energy is large e n o u g h so that the projectile can penetrate into the target nucleus then transfer reactions occur for strongly asymmetric systems or deep-inelastic reactions for m o r e symmetric systems. These reactions provide the only possibility of p o p u l a t i n g n e u t r o n rich nuclei until radioactive nuclear beams b e c o m e available. F o r distant collisions C o u l o m b excitation occurs which allows high-spin states to be p o p u l a t e d by m e a n s of an interaction between the electric fields of the projectile a n d target nuclei. Subsequently the fusion reaction will be discussed in m o r e detail since it provides the most efficient way of p o p u l a t i n g high-spin states in nuclei. In a complete-fusion reaction the incident particle is captured by the target nucleus to f o r m a c o m p o u n d nucleus. T h e kinetic energy in the center of mass system is converted into excitation energy of the c o m p o u n d nucleus. O n e speaks a b o u t a c o m p o u n d - n u c l e u s f o r m a t i o n only if the lifetime of the composite system is large e n o u g h ( > 10~ 2 O s) to reach a t h e r m o d y n a m i c equilibrium before it decays. This m e a n s that energy a n d m a s s are equilibrated. A c o m p o u n d nucleus is formed only if the transferred a n g u l a r m o m e n t u m is below a certain limit, otherwise fast fission occurs. T h e limiting a n g u l a r m o m e n t u m for complete fusion depends on the mass n u m b e r A of the c o m p o u n d nucleus. It has been calculated by C o h e n et al. [13] considering the equilibrium shape of a r o t a t i n g liquid drop, with surface tension included. T h e results are given in Fig. 3, which shows the critical angular m o m e n t u m for which the fission barrier of a ß-stable nucleus

142

Rainer Μ. Lieder

100

0

0

100

A

200

300

Fig. 3. Stability limit for rotating charged liquid masses of angular m o m e n t u m I as function of mass number A (classical estimates).

vanishes (Bf = 0) as a function of the mass number. Below the curve marked Bf = 8 MeV the fission barriers for the rotating ß-stable nuclei are higher than 8 MeV and particle evaporation can successfully compete with fission. The excitation energy of the compound nucleus can be calculated as [7] Eex = Ecm + Q

(1)

Here Ecm is the energy of the projectile in the center-of-mass system and Q the Q-value of the reaction. The maximum angular momentum transferred in the reaction can be estimated in the sharp cut-off approximation, assuming a peripheral collision, as [7]

max

(2)

Here R is the distance of closest approach between the colliding nuclei, μ their reduced mass and Vc the Coulomb barrier. For the reaction 1 5 0 Nd + 3 4 S = 1 8 4 Os at a projectile energy of 158 MeV in the laboratory system, e.g., the completely-fused nucleus acquires an excitation energy of EeJI = 69.4 MeV and the maximum transferred angular momentum is /max = 50 h. This value is smaller than the limiting angular momentum from Fig. 3. Hence the compound nucleus is formed with a large cross section. A more detailed description of the reaction mechanisms can be found in ref. [7],

New Generation of Gamma-Detector Arrays 2. Deexcitation

of highly-excited

compound

143

nuclei

After equilibration the compound nucleus evaporates nucleons until the excitation energy becomes smaller than the separation energy of the nucleon. The decay proceeds predominantly by neutron evaporation unless the c o m p o u n d nucleus lies close to the proton drip line. With a small probability also high-energy giant-dipole γ-rays can be emitted. The neutron evaporation lowers considerably the excitation energy of the nucleus, viz. by its separation and kinetic energies which add up to 8 - 1 0 MeV per neutron on average. However, the mean angular m o m e n t u m taken away by each neutron is only ä 1ft. Depending on the number of emitted particles various final nuclei are produced. When, after « 10~ 1 5 s, the excitation energy of the nucleus becomes less than the neutron separation energy above the Yrast line, the γ-deexcitation starts. In the upper portion of Fig. 4, as an example, the probability of populating the entry states (population distributions) of the final nuclei 1 7 9 ' 1 8 0 O s are shown [14] for the 150 Nd( 3 4 S,xn) reaction at a beam energy of 158 MeV and the γ-ray decay in a deformed nucleus is schematically depicted. The highly-excited final nucleus deexcites at first by the emission of a few high-energy γ-ray transitions of predominantely electric dipole character which on average take away much excitation energy but little angular momentum [15]. O n e talks about statistical transitions, because the spectrum of these transitions has a statistical distribution, due to the high nuclear level density at high excitation energy. If states are reached which lie a few MeV above the Yrast line, the emission of stretched (AI = 2) electric quadrupole transitions can favourably compete with the emission of El transitions [15,16]. They take away the angular m o m e n t u m in the most efficient way in steps of 2 ft until the ground state is reached. However, at this excitation energy the nucleus does not only deexcite along rotational bands but also between them. This behaviour is caused by band mixing, also called rotational damping [16,17]. This process allows the excitation energy of the nucleus further to be reduced towards the Yrast line. It gives rise to a quasi-continuum in the γ-ray spectrum in the energy range of « 0.8 — 1.5 MeV (collective bump) because of the large number of stretched E2 transitions of varying energy and small intensity [16]. If regular band structures within about 1 MeV above the Yrast line are populated then the intensity eventually becomes large enough to allow an identification of the γ-ray transitions, depending on the sensitivity of the γ-detector array. These discrete γ-ray transitions sit in the spectrum on top of a background formed by the statistical decay, the quasicontinuum and the C o m p t o n events as can be seen for 1 8 0 O s in the lower portion of Fig. 4. The ground state of the nucleus is reached after « 10" 9 s and the total number of emitted γ-rays (γ-ray multiplicity My) can be « 30.

C. In-beam γ-spectroscopy techniques The in-beam γ-spectroscopy techniques have been described in detail in textbooks, see e.g. ref. [7,18,19,20], and in publications cited therein. Therefore, it is sufficient to give a summary only here. A detailed study of the properties of nuclear states, e.g. at high spin, requires a high-resolution γ-ray spectroscopy with Ge detectors. The accessible

144

Rainer Μ. Lieder

0.2

OA

0.6

OB

1.0

1.2

1.4

1.6

1.8

Ε (MeV) Fig. 4. Schematic picture of the γ-ray deexcitation of highly-excited nuclei (upper portion) and the resulting γ-ray spectrum measured with an escape-suppressed G e detector (lower portion). The final nuclei have been produced in a 1 5 0 Nd( 3 4 S,xn) reaction at a bombarding energy of 158 MeV. The population distributions (not unfolded) have been measured with a sum-energy and γ-ray multiplicity filter. T h e experiments have been carried out at the VICKSI accelerator of the H M I Berlin with the γ-ray spectrometer OSIRIS.

energy range extends from a few tens of keV to several MeV. In order to establish the complex level scheme of a nucleus it is necessary to place the observed γ-ray transitions and to determine their intensity / γ . This information can be deduced from γ - γ coincidence measurements, since the determination of coincidence relationships between γ-ray transitions forms the basis for their placement in the level scheme. In order to place weak transitions or members of unresolved multiple γ-ray peaks, high-fold coincidence measurements are required. The nuclear states are characterized by their basic eigenvalues and q u a n t u m numbers, such as excitation energy E, spin I and

New Generation of Gamma-Detector Arrays

145

parity π. In order to determine these quantities the features of the γ-ray transitions depopulating these levels have to be measured. They are the transition energy £ , the multipole order L > 1 and the character of the radiation (magnetic or electric) [7], If the features of the final state, populated by the γ-ray are known, then those of the initial state can be determined considering energy conservation £ ; — £ f = Ey and the selection rules |/j — I \ < L < I — 1 , π π = (— \) for electric radiation and π ^ = ( — 1 ) L + 1 for magnetic radiation. In most cases the lowest one or two multipole orders, being in agreement with the selection rules, are allowed for a γ-ray transition, e.g., a mixture of M l and E 2 radiation. i

l

{

L

ί [

The transition energy Ey can be extracted from the location of the corresponding γ-ray peak in the spectrum. The peak area provides information on the intensity of the transition I y . The multipole order and multipole mixture can be deduced from γ-ray angular distribution or γ-γ angular correlation measurements and the multipole character can be obtained from the study of the linear polarization of the γ-rays. It should be mentioned that these values can also be deduced from the measurement of conversion electrons [ 7 , 1 9 , 2 0 ] , The intensity of a γ-transition emitted from an oriented nuclear state has an anisotropic spatial distribution. This is the case for in-beam γ-spectroscopy experiments since the spins of nuclei in excited states, formed in nuclear reactions, are generally aligned in a plane perpendicular to the beam direction [ 7 , 1 9 , 2 0 ] . Correspondingly, the susbstates with quantum number m ( m = — / , . . . , / ) , resulting from the projection of the angular momentum on the beam axis as quantization axis, are not equally populated. The substate population distribution must fulfil the condition P(m) = P( — m) and has been described by a Gaussian [ 7 , 1 9 , 2 0 , 2 1 , 2 2 ] centered around m = 0. The angular distribution of γ-rays is described as [ 7 , 1 9 , 2 3 ] : Π Φ

Ί

)

=

A

0

+

A

P

2

2

( c o s

βγ) +

A

4

P

4

(COS

(3)

ΘΊ)

where P k (cos 0γ) are Legendre polynomials and ΘΊ is the emission angle of the γ-ray with respect to the beam direction. The angular distribution coefficients A / A and A J A depend on the substate population distribution of the initial state, the spins of the initial and final states, the multipole orders and the multipole mixture of the γ-radiation. The study of γ-γ angular correlations requires a coincidence measurement and allows weak transitions or members of unresolved multiplets to be investigated. They could be determined with sufficient statistical accuracy only after the introduction of arrays with many γ-ray detectors, placed under different angles with respect to the beam direction. The theory of directional correlations of γ-radiation emitted from oriented states is well described in the refs. [ 2 4 , 2 5 ] , In many cases a detailed investigation of γ-γ angular correlations is replaced by the observation of directional correlations of γ-rays deexciting oriented states ( D C O ratio method) [26], The D C O ratio is defined as: 2

0

0

\ ν φ R, D C O

—

ι 2

, θ

η

, Φ )

m e y l ,βγ2,φ)

(4)

146

Rainer Μ. Lieder

where ö y l and 0 γ2 are the emission angles of the γ-rays with respect to the beam direction and Φ is the angle between the two planes opened by the emission directions and the beam axis. For linearly polarized γ-rays emitted from oriented states, the electric field vector is parallel (or perpendicular) to the emission plane, defined by the beam direction and the direction of the γ-rays, if they have M l , E2,... (or El, M2,...) character [19,20]. The linear polarization angular distribution W(6 , φ) can be written as an expansion in Legendre and associated Legendre polynomials [20,27,28], The coefficients of the former are the angular distribution coefficients of Eq. 3 and the coefficients of the latter contain the information on the multipole character. The angle between the emission plane and the electric field vector is denoted by φ. The polarization parameter is defined as [27,28]:

K y>

W(ΘΊ, 0°) 4- W(9y, 90°)

^

and takes values in the range — 1 Ρ(0 γ )< + 1 . The sign is determined by the multipole character of the γ-radiation. For pure dipole or quadrupole transitions the absolute value of the polarization parameter is maximum if the γ-rays are emitted at 9y = 90°. The linear polarization of γ-rays is usually measured with a Compton Polarimeter. This is based on the sensitivity of the Compton scattering process to the polarization of the γ-rays, the scattering cross section being larger in the direction perpendicular to the electric field vector than in the direction parallel to it, as described by the Klein-Nishina formula [29]. The basic Compton Polarimeter consists of an active scatterer surrounded by two or more analyser detectors. A useful event for the measurement of linear polarization takes place if the incident γ-ray is Compton scattered in the scatterer detector and the scattered photon is fully absorbed in an analyser detector. In this case, the sum of the energies deposited in the two detectors gives the energy of the γ-ray while the intesity dependence on the azimuthal scattering angle contains the information on the linear polarization. In order to obtain information about the structure of the nuclear levels one has to determine the matrix elements of the nuclear multipole operators. This requires the study of static and dynamic electromagnetic moments. The static electromagnetic moments are given by the diagonal reduced matrix elements of the multipole operators, whereas electromagnetic transitions are characterized by the non-diagonal reduced matrix elements [7,18,30]. In general there is no connection between the diagonal and non-diagonal reduced matrix elements. For specific nuclear model assumptions, however, such connections may exist. Information on dynamic electromagnetic moments is obtained from transition probabilities which requires measurements of lifetimes, branching ratios and mixing ratios. The total transition probability Τ for the decay of a level is obtained from the mean lifetime τ as T = l/τ. If a level is deexcited by several transitions with the partial transition probabilities T{ then the total transition probability is T = Σ T{. The partial transition probabilities can be calculated if the branching ratios are known. The partial γ-transition probabilities ΤJ are obtained after correction for internal electron conversion. The multipole orders and characters of the transitions must be known to calculate the dynamic reduced electromagnetic moments.

New Generation of Gamma-Detector Arrays

147

If a transition is a mixture of different radiations, e.g., M l and E2 radiation, then the multipole mixture must be determined to calculate the transition probabilities for the multipole radiations, e.g., T?(M l)and T}(E2). The mean lifetime τ of a nuclear state can be determined by the study of its exponential decay curve, i.e., the intensity of a transition deexciting this level has to be measured as a function of time. Lifetimes in the range of κ 1 0 " 1 0 to % 10^ 6 s can be measured with the electronic timing technique if Ge detectors are used. Shorter lifetimes can be determined with Doppler-shift methods since nuclei produced in reactions recoil with a certain velocity. The shift in energy of γ-rays, emitted from moving nuclei, can be determined with a Ge detector. To measure lifetimes in the range of « 1 0 " 1 2 to « ~ 9 s the recoil-distance Doppler-shift method ( R D M ) is used. The recoiling nuclei leave the thin target and fly in vacuum with a constant velocity until they are stopped in a plunger placed behind the target at various distances. A decay curve is obtained by measuring the fraction of the γ-ray intensity appearing in the stopped component as function of the flight time between target and stopper calculated from the distance and recoil velocity. Lifetimes in the range of « 1 0 " 1 5 to « 1 0 - 1 1 s can be measured with the Doppler-shift attenuation method (DSAM). The recoiling nuclei are stopped by atomic and nuclear collisions in a thick backing attached to the thin target. If the nuclei decay during the slowing down process then the emitted γ-radiation is energy shifted. This results in a characteristic broadening of the γ-line shape from which the lifetime can be deduced. T o determine the lifetime of a nuclear level one has to correct for all delays occurring in the deexcitation of the highly-excited nucleus. This is especially important for lifetimes smaller than « 1 0 " 1 1 s measured with Doppler-shift methods. These problems can be avoided in γ-γ coincidence experiments which become feasable with large γ-detector arrays. A detailed description of these methods and references can be found in chapter 12 of this book. The determination of static electromagnetic moments requires the measurement of the observable static magnetic dipole moment μ{1) and the observable static electric quadrupole moment Q(I) of a nuclear state. The observable static magnetic dipole moment μ(Ι) is usually expressed by the dimensionless gyromagnetic ratio or g factor as μ{1) = ί / / μ Ν , μ Ν being the nuclear magneton. The observable static electric quadrupole moment Q(I) is defined in the space-fixed frame of reference (laboratory). The electromagnetic moments can be determined by the study of hyperfine interactions with static electromagnetic fields [7,18,19,20], The interaction between the nuclear magnetic moment and a magnetic field Β causes a precession of the nuclear spin around the field direction with the Larmor frequency coL = gß N B/h. The spin rotation can be observed for isomers with lifetimes larger than « 1 0 ~ 8 s by the measurement of perturbed angular distributions or angular correlations. The time dependent perturbed angular distribution (TDPAD), can be described as W{0V t ) = W(0y — co L i) if the magnetic field is perpendicular to the beam-detector plane. In a similar way the electric quadrupole moment can be measured utilizing the hyperfine interaction with an electric field gradient (EFG). The E F G is most commonly supplied by a crystal lattice of non-cubic symmetry. For aligned nuclei only the absolute value of the quadrupole moment can be measured with the T D P A D method. The sign can be determined as well if the recoiling reaction products are polarized before they are implanted in a single

148

Rainer Μ. Lieder

crystal. The recoils can be polarized by letting them fly through a stack of thin foils positioned at a certain angle with respect to the beam direction (tilted multi-foil technique) [7,31], The magnitude and signs of static and dynamic electric multipole (mainly E2) matrix elements can, furthermore, be measured in multiple Coulomb excitation (MCE) experiments [7,20,32,33]. The excitation of high-spin levels of stable nuclei in M C E experiments with heavy ions is the result of a multistep process induced by the Coulomp interaction between the projectile and the target nucleus. With heavy projectiles, e.g. 2 0 8 P b , low-lying collective bands have been excited up to a spin of / = 30, as described in ref. [34], The study of magnetic moments of high-spin states with lifetimes as short as 10 ~ 1 2 s requires very high magnetic fields to obtain a measurable perturbation of the angular distribution or correlation [7], Such a field is produced if the recoiling nuclei are moving with sufficient velocity through a ferromagnetic foil. It results from an ion-solid interaction which increases the density of polarized electrons at the nucleus [35]. The so-called transient field points in the same direction as the external field used to magnetize the ferromagnetic medium and can be as large as several 1000 Tesla, see ref. [35], So far, methods to study discrete γ-rays have been discussed. However, a considerable fraction of an in-beam γ-ray spectrum consists of the quasi-continuum of collective transitions and of statistical transitions {cf. Fig. 4). In order to establish the entry states from which the γ-ray cascades start (the population distributions, shown in Fig. 4) it is necessary to measure the total γ-ray energy ΕΣ released in the nuclear reaction, the number of emitted γ-rays Μ γ as well as the multipole order and the character of the radiation. To extract these quantities for individual γ-ray cascades the continuum part of the spectrum has to be investigated with crystal balls [7], They consist of a large number ( > Μ γ ) of scintillator detectors with a large detection efficiency which cover optimally the total solid angle of 4π. Crystal balls are either operated as stand-alone systems, such as the Heidelberg-Darmstadt crystal ball and the O a k Ridge spin spectrometer [7], or are coupled to high-resolution γ-detector arrays. The HeidelbergDarmstadt crystal ball, being the largest one, consists of 162 Nal(Tl) detectors and has a γ-ray multiplicity and sum-energy resolution of χ 20%. Most of the crystal balls coupled to high-resolution γ-detector arrays (also called sum-energy and γ-ray multiplicity filters) consist of 50-80 B G O scintillators with a γ-ray multiplicity resolution of ä 50% and a sum-energy resolution of « 30%.

II. Performance of gamma-detector arrays The performance of a γ-detector array consisting of escape-suppressed detectors is characterized by several features, viz. the total photopeak efficiency P p h , the energy resolution AEy and the peak-to-total ratio P/T (cf. Sect. I.A). To recognize the relative importance of these features in the detection of γ-rays produced in nuclear reactions it is necessary to define a figure-of-merit for γ-detector arrays. A reasonable figure-ofmerit is the observational limit, being the minimum intensity of a γ-ray transition that

New Generation of Gamma-Detector Arrays

149

can be detected. T h e observational limit is inverse p r o p o r t i o n a l to the sensitivity of a γ-detector array. T h e mechanism in which highly-excited nuclei are p o p u l a t e d a n d d e p o p u l a t e d affect the features of γ-detector arrays ( c f . Sect. I. B). Especially, it has to be taken into a c c o u n t that in heavy-ion induced reactions the nuclei p r o d u c e d fly with a velocity which is a few percent of the velocity of light and that the γ-ray multiplicity is M y χ 30. T h e recoil velocity causes a D o p p l e r broadening of the γ-ray lines d u e to the finite solid angles of the detectors a n d hence a deterioration of the energy resolution. T h e large n u m b e r of γ-rays emitted by the highly-excited nucleus causes with a certain probability the detection of m o r e t h a n one γ-ray in a detector thus reducing the total p h o t o p e a k efficiency a n d the peak-to-total ratio. Subsequently the figure-of-merit a n d the features of γ-detector arrays will be discussed in detail.

A. Figure-of-merit In order to identify a γ-ray peak in a spectrum it must contain a certain n u m b e r of counts to be statistically significant and the p e a k - t o - b a c k g o u n d ratio must be large e n o u g h to distinguish the peak from the statistically fluctuating b a c k g r o u n d . T h e m i n i m u m intensity of a γ-ray transition that can be detected with a γ-detector a r r a y will be deduced for a cascade of Μ γ transitions which occurs in the d e p o p u l a t i o n of a highly-excited nucleus [36]. It is assumed that F detectors of the array are hit by γ-rays of this cascade, i.e., that F-fold coincidences are measured in the experiment. T h e resulting observational limit represents the figure-of-merit for high-spin γ-ray spectroscopy experiments. T h e n u m b e r of counts in a F-dimensional peak can be calculated for a cascade of Μ transitions with an intensity a 0 , normalized to the sum of the intensities of all transitions feeding the g r o u n d state, as

(6) T h e factor 0.76 results since a F-dimensional v o l u m e element determined by the F W H M of the peak is considered. N{*2 is the total n u m b e r of F-fold coincidences obtained by u n p a c k i n g all the measured events with fold ^ F. T h e n u m b e r of F-fold coincidences before u n p a c k i n g are calculated as

(7) where N0 is the n u m b e r of source events in the experiment and P s = Ρ J(P/T) is the total detection efficiency of the array after escape suppression. In order to identify a peak it must stand out f r o m the b a c k g r o u n d . T h e b a c k g r o u n d can be calculated in a simple way, assuming t h a t it is not correlated with the peak events. Indeed, in in-beam γ-spectroscopic experiments a large fraction of the b a c k g r o u n d is u n c o r r e c t e d since it results essentially f r o m the collective transitions in the

150

Rainer Μ. Lieder

quasi-continuum, which are not in coincidence with the cascade. The background can be calculated assuming that the transitions in the cascade have on average an energy separation SEy [6,36,37], Then the background under a F-dimensional peak of width Δ£ γ is approximately

b

V

SEJ

M

,

The background under a peak becomes smaller as F increases. Loosely speaking the reduction of the background can be understood as a dilution. The uncorrelated background is distributed in a space of increasing dimension of size (SE y ) F , whereas the photopeaks are confined to small volume elements of size (Δ£ γ ) Γ . For Μ » F and N ( f ) « iV(bF) the peak-to-background ratio can be calculated as

%

=

aoRo(0.76Rf

(9)

with R

_

SEy Ρ

(10)

being the background-reduction factor of the γ-ray spectrometer. The quantity R is also referred to as resolving power [6,37], The uncorrelated background can also be reduced by other selective devices, providing selection by sum energy, γ-ray multiplicity, recoil detection, etc. To take this into consideration the background-reduction factor Ra has been introduced. To deduce Eq. 9 it has been assumed that the background is completely uncorrelated. If one takes into account the correlated background resulting from Compton contributions as well as statistical and feeding γ-rays, which are in coincidence with the considered cascade, the background-reduction factor becomes fold dependent. This results in an overestimation of the peak-to-background ratio if N^K A better approximation is [38]

with(N^Vb)F_,=aoKo0.76K1. The number of counts calculated according to Eq. 6 corresponds to a case, where a F-l dimensional coincidence condition (gate) is required and the γ-ray peak is observed in the resulting spectrum. However, in many experiments involving highmultiplicity cascades multiple selections can be made thus greatly increasing the number of counts in the final peak. This is, e.g., the case for superdeformed or hyperdeformed bands. The number of peak events for a rotational transition

New Generation of Gamma-Detector Arrays

151

N{pF) obtained by gating on Mb band members in F — 1 dimensions can be calculated for F-fold coincidence data as (Mb-l)(Mb-2)...(Mb-F+l)

KF) = (°·76TJ^) Μ

J,,,«„^

(Λί — 1)(M — 2 ) . . . ( Μ —F + 1)

This corresponds to a summed coincidence spectrum obtained by adding all coincidence spectra resulting from gating on M b members of the rotational band. The observational limit depends on two parameters which are properties of a γ-detector array, viz. the background-reduction factor R and the total photopeak efficiency P p h . To determine the observational limit of a γ-detector array one has to calculate the intensity a c as a function of fold F for these two parameters and to determine the intersection between the two resulting curves. In Fig. 5 two sets of such curves are shown. The variation of aB with F is calculated for various values of the parameter R according to Eq. 9 assuming a peak-to-background ratio of 0.2. The

Fig. 5. The observational limit of a γ-detector array as function of fold for various values of the background-reduction factor R and the total photopeak efficiency P ph , respectively. The numbers used in the calculations are given in the text.

152

Rainer Μ. Lieder

160MeV "CL+"°Pd. SD BAND.

800

1000 Energy (keV)

Fig. 6. Gamma-coincidence spectra of the superdeformed band in 1 4 3 Eu resulting from the analysis of four-fold coincidence data taken with the γ-detector array NORDBALL. No background has been subtracted. For details see text.

New Generation of Gamma-Detector Arrays

153

dependence on the parameter Pph is calculated according to Eq. 12 assuming that gates are set on M b = 10 transitions in a cascade of Μ γ = 30 γ-rays. The peak is considered to contain 100 counts and the total number of source events is taken as N a = 2.6· 10 1 0 . It can be seen in Fig. 5 that the observational limit due to the background reduction factor R improves strongly with increasing fold. However, the statistical limit, which depends on the total photopeak efficiency Pph, reduces with fold. The optimum point for a γ-detector array is where a pair of lines of given values for R and P p h cross. F o r example, assuming R = 12 and P p h = 0.08 one finds an observational limit of 1.5· 1 0 " 5 requiring the measurement of 4-fold coincidence events. In order to demonstrate that the background is reduced with increasing fold, the analysis of four-fold coincidence data [39] taken with the γ-detector array N O R D BALL [40] are shown in Fig. 6. In a study of the superdeformed band in 1 4 3 E u the final nucleus has been produced by bombarding 11 °Pd with 160 MeV 3 7 C1 projectiles. Using the observed 50 million four-fold coincidences, four γ-ray spectra with different gating conditions have been produced. The spectrum displayed in the lowest panel is a total projection of all events. The other spectra result from one-, two- and three-fold gating, respectively. Any combination of the 16 known transitions of the superdeformed band lying between 484.2 and 1325.5 keV has been utilized. The corresponding spectra, shown in Fig. 6, which are not background subtracted, clearly demonstrate the improvement of the peak-to-background ratio. Whereas the superdeformed transitions cannot be seen in the total projection they form prominent peaks in the triple-gated spectrum, with very little background left. For the 733.0 keV line peak-to-background ratios oi(Np/Nb)F = 0.31 ± 0.02,1.03 + 0.05 and 4.4 + 0.5 for F = 2,3 and 4, respectively, have been deduced. The improvement of the peak-to-background ratio with fold clearly demonstrates the importance of measuring high-fold coincidence data in order to study weakly populated bands.

B. Energy resolution The intrinsic γ-ray energy resolution of Ge detectors is typically ΔE'" t r /E y < 0.002 at 1.33 MeV. The resolution is degraded by Doppler effects if the γ-rays are emitted from recoiling nuclei. For heavy-ion induced compound nucleus reactions the recoil velocity has typically a value of a few percent of the velocity of light. Here a value of v/c = 0.025 will be used. The Doppler broadening arises from: 1. the opening angle of the γ-ray detector. 2. the angular spread of the recoils. 3. the velocity (energy) variation of the recoils. The first two effects influence the energy resolution of the detectors the strongest if they are placed at an angle of 6y = 90° with respect to the recoil direction. The last two effects result from particle emission and from slowing down of the recoils in the target. The third effect mainly influences the energy resolution of detectors placed in forward and backward directions.

154

Rainer Μ. Lieder

Values of the half-opening angle of Ge detectors used in present γ-detector arrays range from δ χ 5° to 8°, corresponding to solid angles of üGe = 0.0023 to 0.0051, relative to 4π. This gives a Doppler broadening

(13)

A £ ^ / £ Y = 2-sin0 Y sin 143°) angles, respectively. The distance of the escape-suppressed Ge detectors from the target is 24.8 cm. The detector arrangement can be seen in the photograph of Fig. 26. The features of the G A M M A S P H E R E I array [6] are given in Table 5. It has a total photopeak efficiency of P p h = 0.028 and a background-reduction factor of R = 6.9 for 1.33 MeV γ-rays. The large value of the background-reduction factor is due to the small Doppler broadening in this geometry. G A M M A S P H E R E I has been operative since 1993 at the Lawrence Berkeley Laboratory, USA.

B. Future γ-detector arrays A new generation of γ-detector arrays of unprecedented performance is presently under construction. They will become available in 1996/97, being the arrays

182

Rainer Μ. Lieder

Fig. 26. Photograph of the γ-detector array G A M M A S P H E R E I.

G A M M A S P H E R E (USA) and E U R O B A L L III (Europe), which will be described subsequently.

1. The full GAMMASPHERE

array

The full γ-detector array G A M M A S P H E R E will consist of 110 detector modules [6,37]. A sketch of G A M M A S P H E R E is shown in Fig. 27. In this design the surface of the sphere is covered by a polyhedron consisting of 110 hexagons and 12 pentagons. The detectors are mounted in the hexagon faces and the pentagons, reduced in area, are used for the beam tube and support structures. It is planned that 80 segmented Ge detectors (cf. Sect. III. C) should be used for angles ranging between 37° and 143° with respect to the beam direction, to reduce Doppler broadening. The features of the G A M M A S P H E R E array [6] in this configuration are given in Table 5. It will have a total photopeak efficiency of P p h = 0.101 and a backgroundreduction factor of R = 6.8 for 1.33 MeV γ-rays. The full array is expected to be operational in 1996 but intermediate stages will become available for experiments

New Generation of Gamma-Detector Arrays

183

Fig. 27. Sketch of the γ-detector array G A M M A S P H E R E .

earlier. The array is movable and will initially be sited at the Lawrence Berkeley Laboratory, USA.

2. The EUROBALL

III

array

The γ-detector array E U R O B A L L III will be built by six European countries, viz. Denmark, France, Germany, Italy, Sweden and the U.K., according to a design shown in Fig. 28. It will be based on the E U R O G A M II array and will keep the 24 C L O V E R detectors in two rings around 90° with respect to the beam direction. The backward quadrant will be covered with 15 C L U S T E R detectors (cf. Sect. III. B. 2) and the forward q u a d r a n t with 30 escape-suppressed Ge detectors from G A S P and E U R O G A M . The G e detectors of the latter two types of spectrometers have distances of 43 and 37 cm from the target, respectively. The E U R O B A L L III array will consist of 69 spectrometers using 231 Ge crystals. The features of the E U R O B A L L III array [6,36] are given in Table 5. It will have a total photopeak efficiency of P p h = 0.103 and a background-reduction factor of R — 7.0 for 1.33 MeV γ-arrays. The full array is expected to be operational in 1997. It is movable and will initially be sited at the Laboratori Nazionali di Legnaro (Padova), Italy and subsequently at the Centre de

184

Rainer Μ. Lieder

Fig. 28. Side view of the γ-detector array E U R O B A L L III consisiting of 15 C L U S T E R detectors, 24 C L O V E R detectors and 30 escape-suppressed G e detectors.

Recherches Nucleaires Strasbourg, France. In the array EUROBALL III, as described above, the background-reduction factor is limted by the CLOVER detectors considering that their contribution to the energy resolution is largest, since they are placed around 90° and their Ge crystals have a half opening angle of 4.8° (ιcf. Table 4). A considerable improvement can be achieved if they are replaced by CLOVER detectors of higher granularity like the segmented CLOVER detector (cf. Sect. III. C).

C. Comparison of γ-detector arrays In order to compare the performance of γ-detector arrays it is necessary to calculate the observational limit, which depends on the effective values of the total photopeak efficiency P®" and the background-reduction factor P e f f , determined by the energy resolution AEy and the peak-to-total ratio (P/T) eff . These features are given in Table 5

New Generation of Gamma-Detector Arrays

185

for the existing and presently constructed arrays. They are calculated on the same footing for all arrays using the equations given in Sect. II, assuming a γ-ray energy of Ey = 1.33 MeV, a γ-ray multiplicity of Μ γ = 30, a recoil velocity of 2.5% of the velocity of light and an energy separation of SEy = 70 keV. A neutron multiplicity of M n = 4 is taken into account, making the simplifying assumption that two neutrons each are emitted into the forward quadrant and the 2π region around 90°, respectively and none into the backward quadrant. F o r comparison the features of the first-generation γ-detector array N O R D B A L L [40] are also included in Table 5. The observational limits are deduced from Fig. 5, where it has been assumed that gates are set on 10 transitions in the γ-ray cascade, implying that rotational bands are studied. The fold for which the observational limit is obtained is called optimum fold. A comparison of the features given in Table 5 for the various arrays shows that the observational limit is improved by a factor of « 10 between the first and second generations of γ-detector arrays and that for the future arrays G A M M A S P H E R E and E U R O B A L L III a similar improvement is expected. If a high-multiplicity cascade, such as a hyperdeformed band, is populated with an intensity of a 0 = 10" 5 , then it can be discovered with E U R O B A L L III or G A M M A S P H E R E in a running time of about 10 days for a source event rate of 10 5 /s. The experiment requires that five-fold photo events or eight-fold escape-suppressed events are studied. It is necessary to measure Ä 4· 10 12 unpacked escape-suppressed 8-fold events, corresponding to « 2· 10 1 0 events with a fold of F > 8 before unpacking.

Acknowledgements The material presented in this chapter results from the work of many colleagues. My thanks go to all those whose efforts provided the material upon which the presentation is based. I would like to thank especially W. Gast, A. Georgiev, H.M. Jäger, D. Kutchin, C. Rossi-Alvarez, and S. Utzelmann for clarifying discussions, helpful comments and for support during the preparation of the manuscript. I am particularly obliged to T. Rz^ca-Urban for a critical reading of the article. I am grateful to D. Curien, P. Nolan, C. Rossi-Alvarez, J. Simpson and F.S. Stephens for unpublished material and figures.

References [ 1 ] Morinaga, H. and Gugelot, P. C„ Nucl. Phys. 46 (1963) 210 [ 2 ] Burde, J., Diamond, R. M. and Stephens, F. S., Nucl. Phys. A92 (1967) 306 [3] Twin, P. J., Nolan, P. J., Aryaeinejad, R., Love, D. J. G., Nelson, A. H. and Kirwan, Α., Nucl. Phys. A409 (1983) 343c [4] Sharpey-Schafer, J. F. and Simpson, J., Prog, in Particle and Nuclear Physics 211 (1988) 293 [5] Beausang, C. W., Forbes, S. Α., Fallon, P , Nolan, P. J., Twin, P. J., Mo, J. N , Lisle, J. C„ Bentley, Μ. Α., Simpson, J., Beck, F. Α., Curien, D., de France, G., Duchene, G. and Popescu, D., Nucl. Instr. Meth. in Phys. Research A313 (1992) 37 [6] Nolan, P. J., Beck, F. A. and Fossan, D. B., Ann Rev. of Nucl. and Part. Sei. 45 (1994) 561

186

Rainer Μ. Lieder

[7] Ejiri, Η. and de Voigt, Μ. J. Α., Gamma-Ray and Electron Spectroscopy in Nuclear Physics, Clarendon Press, Oxford (1989) [8] Simpson, J., private communication (1994) [9] Johnson, Α., Ryde, H. and Sztarkier, J., Phys. Lett. 34B (1971) 605 [10] Lieder, R. M. and Ryde, H., Advances in Nuclear Physics 10 (1978) 1 [11] Twin, P., Nyako, Β. M., Nelson, A. H., Simpson, J., Bentley, Μ. Α., Cranmer-Gordon, H. W., Forsyth, P. D., Howe, D., Mokhtar, A. R., Morrison, J. D., Sharpey-Schafer, J. F. and Slettcn, G., Phys. Rev. Lett. 57 (1986) 811 [12] Janssens, R. V. F. and Khoo, T. L„ Annu. Rev. Nucl. Part. Sei. 41 (1991) 321 [13] Cohen, S., Plasil, F. and Swiatecki, W. J., Ann. Phys. 82 (1974) 557 [14] Hebbinghaus, G„ Ph.D. Thesis, Κ FA Jülich Report Jül-2208 (1988) [15] Herskind, Β., Dossing, Τ., Leoni, S., Matsuo, M. and Vigezzi, E., Prog. Part. Nucl. Phys. 28 (1992) 235 [16] Garrett, J. D„ Hagemann, G. B. and Herskind, B„ Ann. Rev. Nucl. Part. Sei. 36 (1986) 419 [17] Lauritzen, B„ Dossing, Τ. and Broglia, R. Α., Nucl. Phys. A457 (1986) 61 [18] Hamilton, W. D. (ed.), The Electromagnetic Interaction in Nuclear Spectroscopy, NorthHolland, Amsterdam (1975) [19] Morinaga, H. and Yamazaki, T., In-Beam Gamma-Ray Spectroscopy, North-Holland, Amsterdam (1976) [20] Pelte, D. and Schwalm, D., in Heavy Ion Collisions, ed. Bock, R., North Holland, Amsterdam (1982) [21] Yamazaki, T„ Nucl. Data A3 (1967) 1 [22] Draper, J. E. and Lieder, R. M„ Nucl. Phys. A141 (1970) 211 [23] Frauenfelder, Η. and Steffen, R. Μ., in Alpha-, Beta- and Gamma-Ray Spectroscopy, ed. Siegbahn, Κ., North Holland, Amsterdam (1995), vol. 2 [24] Steffen, R. M. and Alder, K., in Electromagnetic Interaction in Nuclear Spectroscopy, ed. Hamilton, W. D., North Holland, Amsterdam (1975) [25] Krane, Κ. S„ Steffen, R. Μ. and Wheeler, R. M., Nucl. Data Tables 11 (1973) 351 [26] Krämer-Flecken, Α., Morek, Τ., Lieder, R. Μ., Gast, W., Hebbinghaus, G., Jäger, Η. M. and Urban, W„ Nucl. Instr. Meth. in Phys. Research A275 (1989) 333 [27] Twin, P. J., in Electromagnetic Interaction in Nuclear Spectroscopy, ed. Hamilton, W. D., North-Holland, Amsterdam (1975) [28] Rikovska, J. and Stone, Ν. J., Atomic Data and Nuclear Data Tables 37 (1987) 53 [29] Evans, R. D., The Atomic Nucleus, McGraw-Hill, New York (1955) [30] Bohr, A. and Mottelson, B. R., Nuclear Structure, Benjamin, New York vol. 1 (1969) and vol. 2(1975) [31] Hass, Μ., Dafni, Ε., Bertschat, Η. Η., Broude, C., Davidovsky, F. D., Goldring, G. and Lesser, P. Μ. S., Nucl. Phys. A414 (1984) 316 [32] Aider, Κ. and Winther, Ä„ Ele tromagnetic Excitations, North-Holland, Amsterdam (1975) [33] Cline, D., Ann. Rev. Nucl. Part. Sei. 36 (1986) 681 [34] Piercey, R. B., Hamilton, J. H., Ramayya, Α. V., Emling, H., Fuchs, P., Grosse, E., Schwalm, D., Simon, R. S., Wollersheim, H. J., Evers, D. and Ower, H., Phys. Rev. Lett. 46 (1981) 415 [35] Rud, Ν. and Dybdal, K„ Phys. Scr. 34 (1986) 561 [36] Gerl, J. and Lieder, R. M„ Eds., EUROBALL III Proposal, GSI Darmstadt Report (1992) [37] Deleplanque, Μ. A. and Diamond, R. M., Eds., GAMMASPHERE Proposal, Lawrence Berkeley Laboratory Report 5202 (1988) [38] Twin, P. J„ Nucl. Phys. A557 (1993) 3c [39] Ata?, Α., Piiparinen, M., Herskind, B., Nyberg, J., Sletten, G., de Angelis, G., Forbes, S., Gj^rup, N., Hagemann, G., Ingebretsen, F., Jensen, H., Jerrestam, D., Kusakari, H., Lieder, R. M., Marti, G. V., Mullins, S., Santonocito, D., Schnare, H., Strähle, K., Sugawara, M., Tj0m, P. O., Virtanen, A. and Wadsworth, R., Phys. Rev. Lett. 70 (1993) 1069 [40] Herskind, B„ Nucl. Phys. A447 (1985) 395 [41] Knoll, G. E., Radiation Detection and Measurement, Wiley, New York (1989)

New Generation of Gamma-Detector Arrays

187

[42] Thomas, H. G., Eberth, J., Becker, F., Burkardt, T., Freund, S., Hermkens, U., Mylaeus, T., Skoda, S., Teichert, W., v. d. Werth, Α., v. Brentano, P., Berst, M., Gutknecht, D. and Henck, R„ Nucl. Instr. Meth. in Phys. Research A332 (1993) 215 [43] Lieder, R. M., Jäger, Η., Neskakis, Α., Venkova, Τ. and Michel, C., Nucl. Instr. Meth. in Phys. Research 220 (1984) 363 [44] Michel, C., Emling, Η., Grosse, Ε., Azgui, F., Grein, Η., Wollersheim, Η. J., Gaardh^je, J. J. and Herskind, B„ Nucl. Instr. Meth. in Phys. Research 251 (1986) 119 [45] Baxter, A. M., Khoo, T. L., Bleich, Μ. Ε., Carpenter, Μ. P., Ahmad, I., Janssens, R. V. F., Moore, E. F., Bearden, I. G., Beene, J. R. and Lee, I. Y., Nucl. Instr. Meth. in Phys. Research A317(1992) 101 [46] Beck, F. Α., Byrski, Th., Curien, D., Duchene, G., de France, G., Kharraja, B., Wei, L., Butler, P., Jones, G., Jones, P. and Hannachi, F., in Proc. Workshop on Large Gamma-Ray Detector Arrays, Chalk River, Canada, AECL-10163 (1992) 364 [47] Beck, F. Α., Curien, D., Duchene, G., de France, G. and Wei, L., in Proc. Workshop on Large Gamma-Ray Detector Arrays, Chalk River, Canada, AECL-10163 (1992) 359 [48] Lieder, R. M., Proc. Workshop on Nuclear Structure at High Spin, Bad Honnef, Germany, ed. von Brentano, P., Hiibel, H. and Lieder, R. M., (1989) p. 178 [49] Eberth, J., von Brentano, P., Teichert, W., Mylaeus, T., Lieder, R. M., Gast, W., Hebbinghaus, G., Jäger, Η., Maier, Κ. Η., Grawe, Η., Kluge, Η., Schwalm, D., Gerl, J., Hübel, Η., Henck, R. and Gutknecht, D , Nucl. Phys. A520 (1990) 669c [50] Eberth, J., von Brentano, P., Teichert, W., Thomas, H. G., v. d. Werth, Α., Lieder, R. M., Jäger, H., Kämmerling, H., Kutchin, D., Maier, Κ. H., Berst, M., Gutknecht, D. and Henck, R„ Prog. Part. Nucl. Phys. 28 (1992) 495 [51] Lieder, R. M., Proc. XI Int. School on Nuclear Physics, Neutron Physics and Nuclear Energy, Varna, Bulgaria, ed. Andrejtscheff, W. and Elenkov, D., (1993) p. 152; Eberth, J., Phys. Bl. 49 (1993) 1016 [52] Kutchin, D„ Thesis, KFA Jülich (1994) [53] Brun, R., Bruyant, F., Maire, M., McPherson, A. C. and Zanari, P., C E R N Report DD/EE/84-1 (1986) [54] Kämmerling, Η. and Jansen, P., Report KFA Jülich, Jül-2849 (1993) [55] Gysin, L„ Lieder, R. Μ. and Marti, G„ KFA Annual Report, Jül-2590 (1992) 106 [56] Eberth, J., Proc. Conf. on Physics from Large y-Ray Detector Arrays, Berkeley, USA LBL-35687, CONF-940888 (1994) p. 160 [57] Garcia-Raffi, L. M. and Tain, J. L., private communication (1994) [58] Butler, P. Α., Carr, P. E., Gadeken, L. L., James, A. N., Nolan, P. J., Sharpey-Schafer, J. F., Twin, P. J. and Viggars, D. Α., Nucl. Instr. Meth. in Phys. Research 108 (1973) 497 [59] Protic, D. and Riepe, G„ IEEE Trans. Nucl. Sei. 32 (1985) 553 [60] Berthold, Α., Degenhardt, J., Mourikis, S., Schmitz, J. R., Schülke, W., Schulte-Schrepping, H., Wohlert, F., Hamacher, Α., Protic, D. and Riepe, G., Nucl. Instr. Meth. in Phys. Research A320(1992) 375 [61] Machiavelli, A. O., Lee, I. Y., Cederwall, B., Clark, R. M. Deleplanque, Μ. Α., Diamond, R. M., Fallon, P. and Stephens, F. S., Proc. Conf. on Physics from Large y-Ray Detector Arrays, Berkeley, USA LBL-35687, CONF-940888 (1994) p. 149 [62] Gerl, J., Vetter, Κ., Elze, Th. W., Kroll, Th. and Xie, H., Proc. Conf. on Physics from Large y-Ray Detector Arrays, Berkeley, USA LBL-35687, CONF-940888 (1994) p. 159 [63] Bazzacco, D., Brandolini, F., Buscemi, Α., Lunardi, S., Nardelli, G., Nebbia, G., Pavan, P., Rossi Alvarez, C., Signorini, C., Soramel, F., Zanon, R., de Poli, M., Favaron, P., Maron, G., Montagnoli, G., Prete, G., Spolaore, P., Stefanini, A. M., Vedovato, G., Bortignon, F., Lo Bianco, G. and Sala, P., GASP Proposal, Padova (1990); Bazzacco, D., in Proc. Workshop on Large Gamma-Ray Detector Arrays, Chalk River, Canada, AECL-10163 (1992) 376 [64] Spolaore, P., Larson, J. D., Signorini, C., Beghini, S., Zhu, X. and Si, H., Nucl. Instr. Meth. in Phys. Research A238 (1985) 381 [65] Ender, Ch., Beck, F. Α., Ring, Ch., Wendling, L., Richard, Α., Aleonard, Μ. M., Lazarus, I.,

188

[66] [67] [68] [69] [70]

Rainer Μ. Lieder Puckneil, V., Cresswell, J. R., Nolan, P. J., Holm, Α., Kluge, Η., Maier, Κ. Η., Ziem, P., Gast, W. and Kossionides, S„ IEEE Trans, on Nucl. Sei. 37 (1990) 326 Goulding, F. S., Landis, D. Α., Madden, N., Maier, M. and Yaver, H., IEEE Trans, on Nucl. Sei. 41 (1994) 1140 Goulding, F. S. and Landis, D. Α., IEEE Trans, on Nucl. Sei. 41 (1994) 1145 Alexander, J., Beck, F. Α., Ender, Ch., Lazarus, I., McPherson, G. M., Owen, E. C. G., Pouxe, J., Richard, A. and Ring, Ch., IEEE Trans, on Nucl. Sei. 39 (1992) 886 Georgiev, Α., Gast, W. and Lieder, R. M„ IEEE Trans, on Nucl. Sei. 41 (1994) 1116 Georgiev, A. and Gast, W„ IEEE Trans, on Nucl. Sei. 40 (1993) 770

6

Modern Electron and Positron Spectrometers —Selected Aspects of a Broad Field

H. Bokemeyer,

J. van Klinken

and P.

Salabura

Table of Contents I. II.

Introduction 189 Aspects of electron and positron spectroscopy 191 A. Interaction with matter 191 B. Kinematic shifts and Doppler corrections 193 C. Sources contributing in the lepton channel 196 III. Electrons and positrons moving in magnetic fields 199 A. Toroidal fields 200 B. Solenoid fields 201 IV. Selected instrumental solutions 203 A. Transport field spectrometers 203 1. Solenoid spectrometers for conversion electrons 204 2. Solenoids for positron and pair spectrometry 207 B. Dispersive spectrometer devices 215 1. Orange-type spectrometer systems 215 2. Mini-orange spectrometer family 220 C. Trajectory spectrometry from below 100 MeV till beyond 1 GeV V. Remarks 230 References 232

226

I. Introduction Studies of the emission of electrons and positrons, separately or in pairs, are embedded in a broad range of disciplines. Ordered with increasing energy, electrons are observed in chemistry and atomic physics at thermal energies from meV to eVs and as Auger electrons from eVs to 100 keV. In ion-atom collisions electrons are produced and studied with energies up to MeVs. In nuclear transitions electrons are emitted with energies from keVs to MeVs by internal conversion, electrons or positrons by ß-decay and electron-positron pairs by internal or external pair creation. At energies below 100 MeV till 10 GeV Dalitz and dilepton decays of mesons and heavy leptons occur and at still higher energies electron-positron pairs are emitted in the Drell-Yan process of quark-antiquark annihilation. This list of examples covering an extended energy range as well as the diversity of the underlying processes demonstrate, that the

190

Η. Bokemeyer, J. van Klinken and P. Salabura

instrumental techniques and applications for electrons and positrons must be of great variety. In what follows we will select some instrumental solutions applied to in-beam spectroscopy of conversion electrons and discuss some techniques used in electron and positron spectroscopy of quasi-atoms. As an example we will present a search for resonant Bhabha scattering at the low energy of 1.8 MeV and describe a project on dilepton spectroscopy at higher relativistic energies after heavy-ion collisions as a future development. By means of these examples the reader is informed about the scope of modern experimentation at energies from 100 keV up to 1 GeV. N o doubt this view on a broad and vivid field will remain incomplete. We treasure electrons and positrons as our best known system of particle and antiparticle: stable, singly charged, s p i n - | leptons of low mass interacting by electroweak coupling. In the present energy regime its dominant interaction is by electromagnetic forces. This is most favourably used in nuclear and atomic spectroscopy where the electron competes with X-ray and γ-ray emission and carries additional spectroscopic information [1], In heavy-ion collisions quasi-atoms [2] are formed transiently during the short collision times of 10" 2 1 s. The p r o m p t emission of > 0.068 GeV Z 1 / 3 the radiative energy loss becomes ( ^ Λ ~ 1.39· 10" 3 ^ - £ 0 ^ - - l n ( 1 8 3 Z ~ 1 / 3 ) \ p d x j „d g A

(1) pX0

See [9] for screening effects. The range distributions are characterized by the radiation length X 0 given by: — ~ 1 . 3 9 - 1 0 " 3 — Z ( Z + l)-ln(183Z"1/3) g A

(2)

Plural and diffuse scattering are mostly caused by the C o u l o m b interaction with atomic nuclei. They are in general hard to describe in a rigorous and complete way. F o r large numbers of scatterings ( N > 10) and for small deflection angles, the distortion of the primary direction can be described by a Gaussian distribution the mean value being zero. With this distribution projected on a plane, the rms-width obtained after passing

192

Η. Bokemeyer, J. van Klinken and P. Salabura

through a material layer of thickness χ is given by 13.6 MeV

0o =

1 1 6

M

e

V Z

v

^

[ i + 0.0038 l n ( x / * 0 ) ]

(3)

within the range 1 0 " 3 A'Q < χ < IOA'Q. For 500-keV electrons passing through gold layers of a few mg/cm 2 the width θ0 is of the order of 10°. For 1-MeV electrons passing through thin mylar foils of 2.5 μg/cm 2 the width still amounts to 1.5°, but decreases to 0.3° when the electron energy reaches 10 MeV. It is obvious that this angular straggling renders tracking techniques in magnetic fields inpracticable when the electron energy is less than 10 MeV. Below the critical value Ec the electron or positron energy can be determined with often excellent energy resolution by low-Z detectors. In view of background reduction their thickness should not exceed the maximum range of interest. U p o n close scrutiny the thickness of a full-energy detector may even be slightly less than the range [15]. The response of these detectors to electrons differs strongly from the response to γ-rays, the latter being predominantly characterized by large C o m p t o n distributions-very large in case of low-Z detectors - and consequently by small peak-to-total ratios. Electrons and positrons also produce low-energy tails in spectra taken with solid-state detectors, but these tails are relatively small. They are due to outscattering and to Bremsstrahlung losses and a m o u n t to no more than e.g. 18% for perpendicular incidence of electrons in Si at 1 MeV. Outscattering occurs when the incident electron is back-scattered or when it leaves the detector at the edges by multiple C o u l o m b scattering. The amount of backscatterring is nearly constant below 1 MeV, but decreases for higher energies. It is always minimal for perpendicular incidence but strongly increases at oblique angles. The contribution to outscattering is about 10% higher for electrons than for positrons. The tail due to outscattering is flat and extends down to zero energy. The Bremsstrahlung losses impair the detector response in a different way. Since the cross section for Bremsstrahlung is proportional to the inverse frequency, its main effect starts immediately at the low-energy side of the full-energy peak. The detector response to positrons in addition shows a characteristic high-energy tail, which is caused by partial summation of 511-keV annihilation quanta from positronium mostly formed by the incident positron when it is stopped. For typical Si(Li) detectors the tail amounts to between 5 and 10% of the total response. With life-times of the positronium of about 1 0 " 1 0 s this annihilation summation cannot be distinguished electronically from the stopping and the charge collection processes in the detector. At energies beyond Ec when Bremsstrahlung dominates, the energy loss proceeds by electromagnetic showering via cascades of γ-rays, electrons and electron-positron pairs [10, 11], The evolution of the showers strongly depends on the atomic number of the stopping material. The process can be simulated by M o n t e Carlo techniques with the longitudinal energy loss parameterized by (4)

Modern Electron and Positron Spectrometers

193

Here δ equals the dimensionless maximum material thickness i m a x where the number of generated particles reaches its maximum: in units of X0 t max ~ In (E0/Ec) — t0; the offset t0 is around 1.1 for electrons and 0.3 for photons. Γ stands for the Γ-function with the argument Σ ^

(20)

'rf + r +(z — ζ,·) — 2 rtr cos φ

For less symmetric fields we recommend the use of commercially available threedimensional computer codes, like OPERA-3d [27], to describe the transport features. Starting from the Lorentz equation one obtains for the equations of motion: d2r 2

dt

ίδφ\2 r — \dt J

e

=

δφδ{τΑΛ

ymec dt

dr

d2z e δφ δΑφ e δφ 22 = + — T 7 r - ^ = - — r - £ B dt ymec dt δζ ym^c dt e

δ( dt\

dt J

^

ymec dt

e

—=

e

ymec

δφ ν — Βz dt (21)

r

δ

2

2ymec dt

The integration of these equations usually starts with the ^-component [28].

202

Η. Bokemeyer, J. van Klinken and P. Salabura

Homogeneous solenoid fields with Bz = const and Br = Βφ = 0 allow a complete analytical description of the motion of electrons. The path becomes a helix: e · / , , s = V j P b sm( - φ ο) = |e| c ζ = 2—ρΒ(φ-φ0)οοίθ0 |e| r

Φ-Φο=

=

sin0 o sin (φ - φ 0) ωΒ cos 0

(22)

ωΒ

-ωΒί/2.

Here the electron is emitted with velocity v, phase φ0 and emission angle θ0 (with respect to z) at ζ = 0, r = 0. In projection along the axis the helix shows as a circle with radius:

Pb =

^bsinθ°=0333 ϊ ^ βsinθ°= °·171 Tcm ~f"sin00

centered at r = pB, φ = φ0. Along this circle the electron moves with constant velocity v± = ωΒρΒ defining a period Tp = 2π/| ω Β |. Electrons and positrons spiral with opposite signs - electrons right-handed, positrons left-handed - as expressed by the slope of their helical path: d

=

_

dz

tang0 |e|

ωΒ

=

2pB

2DCOS0 o '

Solenoids for positrons use this difference in the spiral sense to discriminate against electrons by helical baffles. Along the z-direction the longitudinal velocity is υ, = vcos θ0. The helix intersects the z-axis periodically with pitch z p : 2πυ0 cos0 o Tcm ρ z p = 2H P B cot θ 0 = - Ί _ Γ - = 2.1 — - cos θ0 = 1.07 T c m ν Ζ Ξ ΐ

cos θο = 7 0

. 1 0 - χ, _T*_

(25) Ev ^

^

The spiraling movement in inhomogenous fields can be described in adiabatic approximation [8] with Β nearly constant during one particle rotation. In that case the gyration center G follows the field in such a way that the electron returns to the field line at which it has been emitted. This is the basic transport feature of solenoids. In some detail the adiabatic transformation scales the shape of the helix between two positions 1 and 2 with different Bz values by: 'sinflA2 J p j S f sin θ2)

\ρΒΛ)

=

Biii

(26)

Bzl

The adiabatic approximation proves an excellent tool for design purposes. For more details, and in particular for the transformation of the phase, numerical integrations of Eq. (21) have to be performed with e.g. the G E A N T code [16],

(23)

Modern Electron and Positron Spectrometers

203

F r o m Eq. (26) the m i r r o r relation sin 2 0o,min = ^z,o/^2,maxf°^ ows with a lower cutoff angle 0 O m i n . Trajectories starting with an emission angle larger t h a n ö0,min reflected when the field increases to a m a x i m u m value Bzmax. This reflexion in a rotationally symmetric field is superimposed by a r o t a t i o n of the gyration center G a r o u n d the symmetry axis. Approximately: dΦα df

ωΒρΒ Bz(0,z)

dB 2 dz

(27)

T h e reflexion occurs simultanously for electrons a n d positrons. Although the r o t a t i o n of the gyration center G of their paths proceed in opposite directions this feature can hardly be used for their separation. The r o t a t i o n is a special case of the gradient shift [ 8 ] for longitudinally i n h o m o g e n o u s fields. In some less symmetric cases it can be used to distinguish between the two charge states of a particle. As a practical example of such a m o r e general i n h o m o g e n o u s field we choose a bent solenoid with a l/r-field distribution. It resembles a sector of a torus with curvature radius r. W h e n emitted with velocity v 0 at an angle 9 0 (relative to the central field line) the e q u a t i o n s of m o t i o n for an electron in this case can be solved in the adiabatic a p p r o x i m a t i o n . Again the electrons spiral with the gyration center G following the field direction. Accordingly, the velocity v(t) of the particle along its p a t h can be decomposed into the velocity of the center vG(t) with vG {t = 0) = v0 cos >90 a n d into the circular velocity ? x ( t ) according the L o r e n t z force so that u(t) = v G (t) + ν λ ( ή . This decomposition leads to the e q u a t i o n of m o t i o n for the gyration center [ 8 , 3 4 ] : du G (t)

^ 2 V B x B = vG(t) χ ώΒ — ν\

(28)

T h e second term is the gradient shift again. T h e position of G at the time t follows [34] f r o m the integration of Eq. (28). Perpendicular to Β a n d V Β we observe a shift: zG(i)= — + c o s coBr 2

2

5

0

) .

(29)

Notice that this shift of the gyration center is the result of the cycloidal m o v e m e n t of an electron in a 1/r-field in a plane perpendicular to the field lines. Bent solenoid spectrometers m a k e use of this shift to obtain a separation of electrons a n d positrons as will be described in the next section.

IV. Selected Instrumental Solutions A. Transport field spectrometers In-beam spectroscopy of electrons and positrons with energies f r o m 100 keV to 1 M e V a n d beyond often requires high resolution in energy and time, as well as high efficiency a n d rigorous suppression of b a c k g r o u n d . T h e use of pulse-height analysis ( P H A )

204

Η. Bokemeyer, J. van Klinken and P. Salabura

benefits f r o m the availability of l o w - Z semiconducting a n d scintillation detectors. H i g h e r - Z detectors with larger stopping powers are sometimes employed for higher energies u p to a n d b e y o n d 10 MeV. However, without further precautions, these otherwise excellent devices for electron a n d positron detection in most cases accumulate an overwhelming b a c k g r o u n d f r o m X-rays, γ-rays a n d from rc and zp(E) > z c can reach the detector. At higher energies the transmission is reduced when the size of the projected trajectory exceeds the radius rd of the detector surface. T h e p l a n a r Si(Li) c o u n t e r is usually m o u n t e d with its surface perpendicular to the solenoid axis, a l t h o u g h we will also describe geometries with longitudinally extended cyclindrical detectors. T h e acceptance of a detector in the solenoid field is easily derived f r o m the geometric overlap of the projected orbit on the detector surface [37]. The t r a n s p o r t efficiency of the one-arm solenoid can be enlarged further if a m a g n e t ic mirror is implemented by raising the field Bz b e y o n d the target value on the o n e side of the solenoid so far not used for electron detection. N o t e that for the same reasons a n excessive increase of the field between target a n d detector will cause a reduction of the efficiency when this field starts to act in p a r t as a mirror. T h e device of ref. 35 offered a nearly flat detection efficiency of 5 % in the energy range from a b o u t 200 keV to 3 MeV. A thin aluminized mylar foil protects the Si(Li) detector against light t h a t m a y be emitted by excited ions and a t o m s f r o m the beam, against rest gas in the v a c u u m chamber a n d against c' = c + V,nl). As a consequence the mass eigenvalues and accordingly Am2, will change while the mixing angle obtained from the diagonalization of the Hamiltonian matrix remains unaffected and identical to its vacuum value (9V. Contrary to m u o n neutrinos, electron neutrinos, however, also interact in a charged current scattering process with electrons. This additional interaction adds an interaction potential F cc (v e e) to only one of the diagonal terms of the Hamiltonian (a' - > a ' + Fcc(vce)) and implies a change of the mixing angle. The relation between the mixing angle in matter, 6>m, and the mixing angle in vacuum 6>v is given by:

sin 2 0

sin20„

=

(4) cos20v

2

+ sin 2 0 v

L 0 denotes the interaction length for charged current (vee) scattering and depends on the electron density pe of the traversed material: L 0 oc 1 /pe. Equation (4) shows that for c o s 2 0 v = / OSC /L 0 the oscillation amplitude in matter reaches a maximum. Thus very small or even not observable vacuum oscillation amplitudes can resonantly be enhanced when neutrinos pass through matter. The implications of this effect in relation to the solar neutrino problem were first discussed by Mikheyev, Smirnov and Wolfenstein (MSW effect) [15].

C. Physics with low energy neutrinos 1. Neutrino oscillation experiments

at nuclear

reactors

With only electron antineutrinos being in nuclear fission processes oscillation experiments at nuclear reactors search for a possible disappearance of the primary ve. Searching for the appearance of muon or tau-neutrinos is not possible since the low energies of reactor neutrinos do not allow the production of the charged leptonic partners (μ or τ) which could identify the incident neutrino flavour. The existence of neutrino oscillations at nuclear reactors can thus manifest itself only in a modulation of the intensity and the shape of the primary reactor v e -spectrum. Detectors are set u p at different distances from the reactor core and, usually by means of the inverse neutron decay reaction v e p - » n e + , the integral neutrino flux and also the neutrino energy spectrum at the detector site is measured.

278

Jean-Luc Vuilleumier and Gabriele Zacek

Two approaches are followed up when analyzing the data in terms of neutrino oscillations. O n e analysis method is based on a comparison of the neutrino spectra or the neutrino yields measured at different distances from the core of the nuclear reactor. It is tested whether possible deviations in the shape or normalization of the individual spectra can be attributed to neutrino oscillations. The results of this analysis are essentially free from the precise knowledge of the reactor neutrino spectrum, the detection efficiencies and the reaction cross section. Since, however, no information about the absolute neutrino flux is used the sensitivity to sin 2 2 Θ ν in the region of large Am2 is

Fig. 2. Limits on oscillation parameters from experiments at nuclear reactors (Gösgen, Kurtchatov). The regions to the right of the curves are excluded with 90% confidence. Limits from accelerator experiments are shown for comparison [16]'

N e u t r i n o Detectors

279

lost. A second method includes the reactor neutrino spectrum in the analysis. It thereby depends on the knowledge of the absolute neutrino flux normalization and detection efficiencies, however, retains its sensitivity to the mixing angle for large values of the mass parameter. Results from neutrino oscillation experiments are usually presented in the form of 2-dimensional exclusion plots showing regions of (Am 2 , sin 2 6>v) parameter pairs which are not compatible with the measured data. Figure 2 summarizes the present limits obtained from neutrino oscillation experiments at nuclear reactors. Parameter pairs lying to the right of the curves are excluded with 9 0 % confidence. U p to now no indication for the existence of neutrino oscillations has been observed. In order to improve the sensitivity of reactor experiments towards smaller values of Am 2 , the distances have to be increased. Sizes must be scaled up, to compensate for the lower neutrino fluxes. Experiments have then to solve the difficult problem of maintaining a low background level in much larger detectors. F o r reactor experiments background sources essentially originate from cosmic ray induced events and from events caused by the natural radioactivity of the detector construction materials. Background correlated with the reactor operation can be suppressed in general by appropriate shielding methods and accidentally induced background events are measured during the periodically scheduled reactor off-periods necessary to exchange part of the fuel elements. Present neutrino detectors located at a distance of about 60m from the core of a power reactor have reached a signal to background rate of about 1 to 1. By improving the shielding conditions against cosmic radiation the aim [17] is to obtain in a 1000t neutrino detector and at a distance of 13 km a signal to background ratio of 8 to 1 and hence to lower the sensitivity for Am 2 down to about 1 0 " 4 eV 2 . In a preliminary step, experiments with detectors of order 10 tons, at distances a r o u n d one kilometer from a reactor, are being considered.

2. Solar neutrino

experiments

Three different reactions are employed in presently operating solar neutrino experiments to measure the flux of incident electron neutrinos: a) ve +

37

Cl->

37

Ar + e "

b) ve + e " ->v e + e " c) ve +

71

Ga->

71

Ge + e"

Reaction a) is used by the pioneering experiment of R. Davis [ 18] which started data taking in 1970. The result of this experiment initiated the discussion of the famous solar neutrino problem, i.e. the fact that the neutrino flux observed on earth fell short of the theoretically predicted one by a factor of about three. A similar result was also obtained by the Kamiokande group [19] which detected solar neutrinos with the help of reaction b) and found only about half of the rate expected according to current calculations. O n e attractive solution to explain these solar neutrino deficits are neutrino oscillations in combination with the M S W effect where the relatively large discrepancy between

280

Jean-Luc Vuilleumier and Gabriele Zacek

observed and calculated solar neutrino fluxes is attributed to the fact that small vacuum oscillations are enhanced when electron neutrinos traverse matter on their way to the earth (sect. II.B.2). However, both experiments have high detection thresholds (£ t h r = 0.814 MeV f o r 3 7C1; £ t h r ~ 7MeV for ve) and are exclusively sensitive to the high energy tail of the solar neutrino spectrum where solar model predictions are uncertain because of their strong dependence on solar parameters. Results from 'Gallium' experiments (reaction c)) can be compared more reliably with solar model calculations, since due to the lower detection threshold (£ t h r = 0.233 MeV for 7 1 G a ) they are sensitive to the well known flux of pp-neutrinos. For 'Gallium' experiments it is predicted that about 60% of the neutrino captures are due to ppneutrinos. First preliminary results of the European Gallium experiment GALLEX [22] and also the new data from the American Russian collaboration SAGE [23], indicate that at least the expected flux of pp-neutrinos has been seen. Still, the observed neutrino fluxes are lower than one would have expected by adding up all contributions of detectable solar neutrino branches. Since, however, also in the case of the 'Gallium' experiments an appreciable amount of the flux has to be attributed to neutrinos of high energies, it might be too early to explain the problem of the missing solar neutrinos solely by neutrino oscillations. In any case one has to keep in mind that the neutrino capture cross sections are extremely small (σ oc 0(1O~ 4 5 cm 2 )) and therefore solar neutrino experiments deal with event samples of quite low statistics. According to the neutrino flux predicted by the standard solar model e.g. a rate of 0.2 events per day is expected for the 133 t 37 C1 detector of R. Davis. Systematic uncertainties which arise e.g. from the knowledge of the neutrino capture cross section and extraction efficiency ('hot chemistry') can probably only be reduced by calibrating the detectors directly with neutrinos. F o r this purpose a measurement with a 1 MCi 5 1 Cr source is discussed in connection with the 'Gallium' experiments. The activity such a source will induce a signal in the detector which exceeds the solar model predictions by a factor of 7 [22,24],

3. The magnetic moment of the neutrino The neutrino magnetic moment matrix μπ,(1,Γ = e, μ,τ), like the mass matrix mu,, is fundamental and its experimental study may provide insight on new physics. In the standard model, neutrinos are massless Dirac particles and have vanishing magnetic moments. In the simplest extension neutrinos are still Dirac particles but may have masses and acquire a magnetic moment through radiative corrections [25,26] 3G F em,

(5)

in Bohr magneton units, much too small however to be of any significance in lab experiments or in astrophysics. In left-right symmetric models the magnetic moments can be more important, of order 1 0 " 1 4 for the ve. This is still small for presently planned

Neutrino Detectors

281

experiments. But models predicting moments up to 1 0 " 1 0 have been built (see for example [27-32]. These models incorporate new particles, for instance they have a richer Higgs sector. Neutrinos may be of Dirac or Majorana type, with the restriction, in the Majorana case, that μη. is anti-symmetric and only transition moments connecting different flavors are allowed, but not static moments. Magnetic moments will give the neutrinos electromagnetic interaction, allowing scattering from the active left handed v lf states into the sterile ν ΓΛ states. Scattering cross-sections for both neutrino-baryon and neutrino-electron will be affected, which makes it possible to find evidence for magnetic moments in lab experiments. The contribution, compared to that of the usual weak interaction, increases with decreasing neutrino energy [33]. But at low energy the baryon recoil is very small and difficult to measure. The electron recoil energy is much larger and the best way of looking for the magnetic moment of the ve seems to be the detailed study of vce_->vce~

and

vce"->vee"

(6)

scattering at low energy. M o r e generally these reactions are fundamental and their study provides information on basic features of the weak interaction. Both charged ( C C ) and neutral currents ( N C ) are involved. They are expected to interfere if the N C and C C final state neutrinos are identical, as assumed in the standard model [34], A measurement of the differential cross section of either reaction allows, in principle, to determine the Weinberg angle sin2 0W and to observe the interference which is expected to be destructive for reasonable values of sin2 0W. Practically, however v e e ~ —• v e e ~ only has a good sensitivity to both effects, while v e e~ -> v c e~ essentially probes the interference only. The differential cross section is given by [33,34]: άσ dΤ

Τ Y

G£me 2π

(0V +

X +

0 A )

2

+ ( 0 V

+

* - . 9 A )

2

(

1

-

J

I

m Τ + ( 0 A - ( 0 V

+

* )

2

) ^ 2 -

where Τ is the electron kinetic recoil energy, £ v the neutrino energy; in the standard model one has: ^ • 0 + -1, gv = 2sin W

gA = iί

1

for ^

v _t

Here VTIÄU5 I

is the effective magnetic moment; χ is calculated from the neutrino form factors, for

282

Jean-Luc Vuilleumier and Gabriele Zacek

Dirac neutrinos it is related to the square charge radius of the neutrino w

χ = ——^sin 2 0 w

for

ve,

x - > —χ

for ve

(8)

The terms on the first line are due essentially to W and Ζ exchange. Their contribution to the total cross-section increases linearly with Ev. The term on the second line comes from the magnetic moment interaction, and its contribution increases only logarithmically. To observe this term one must use low energy neutrinos. However, the most spectacular consequence of magnetic moments will be the precession of neutrinos from left handed v1L states to right handed v,,R states in the presence of a transverse magnetic field B±. In vacuum, assuming for simplification two neutrino states 1 and Γ diagonal in the mass, the evolution is given by:

(9) Here Ε is the neutrino energy, and Am2 = ra2R — mfL is the difference of the squared masses. In case of equal masses, for instance if the neutrinos are Dirac with 1 = Γ, the precession frequency is |μ„,| |ß ± |. This is too small to have any effect in terrestrial experiments, but may be important in astrophysical systems. In fact, solar neutrino astronomy is at the origin of the recent interest in the magnetic moment of the neutrino. Such a moment might be responsible for the observed low flux of 8 Β ve from the sun in the 3 7 C1[35,18] (also see ref. [16]) and Kamiokande experiments [19]. These neutrinos have relatively large energies, several MeV. Voloshin, Vysotskii and O k u n [36] have pointed out that the sun, in solar active years, has a nearly toroidal magnetic field above and below the equator, in the 2· 10 1 0 cm deep convective zone. The strength, during solar active years, is of order a few kG. If some με1 is of order 10 1 10 " 1 1 , the 8 Β ve, which are produced very near the center of the sun, may flip their spin on their way out when crossing the magnetic field, and become sterile. One then expects the solar 8 B flux to be modulated with the 11 years solar cycle, which is in fact suggested by the 37 C1 data, although not observed by Kamiokande. To be precise, in addition to the magnetic moments, a complete treatment of neutrino propagation must take into account the interaction with matter and the possibility of non-diagonal masses [37,38]. Suppression of the precession, or complete conversion from one state to another, can then occur. The magnetic field distribution in the sun's interior is also important. It is presently poorly known, and assumptions must be made. Many scenarios are possible, depending on the choice of parameters [39,40]. In ref. [41] it is shown that one can, assuming magnetic moments and diagonal masses only, reconcile the fluxes and time dependence of the 37 C1 and K a m i o k a n d e experiments, as well as the flux of low energy neutrinos, well below expectations, reported by the 7 1 G a G A L L E X and SAGE experiments. Also, one may include three families of fermions in the calculations, instead of two, as is done usually to simplify [42]. Astrophysical observations suggest that, with moments of order 1 0 ~ 1 0 - 1 0 ~ u , neutrinos would be of the M a j o r a n a type. If neutrinos were Dirac particles, then the

Neutrino Detectors

283

observed duration (several seconds) of the SN1987A neutrinos burst implies much smaller values: μ„. < 1-20-10" 1 3 [43,44,45]. These limits assume that right-handed neutrinos are sterile and can escape from the supernova, and in any case do not apply to Majorana neutrinos. Limits from stellar cooling apply to both Dirac and Majorana neutrinos, but are less stringent: μ„, < 3-10" 1 2 - 1 0 " 1 1 [33,46,47,48,49], This is not in strong conflict with magnetic moments of the strength considered, in particular if one takes into account the model dependence of the astrophysical calculations. Turning around the argument, one can say that, should the neutrino have a large magnetic moment, it would play an important role in stellar physics. It therefore appears important to look for the neutrino magnetic moment in a lab experiment. The best results have been obtained with low energy neutrino sources, at beam dumps and, in particular, at nuclear reactors in the study of ve — e scattering.

III. The detection of low energy neutrinos A. Charged current neutrino capture reactions 1. The inverse ß-decay reaction of the neutron The inverse ß-decay reaction of neutron: ve + ρ -> e + + n has the important advantage that by measuring the energy E e of the final state positron the energy £ v of the incident neutrino can directly be inferred. Neglecting nuclear recoils the relation between the kinetic energy of the positron and the neutrino energy is given by: Ev = Ee + 1.804 MeV, where the neutrino detection threshold of 1.804 MeV is just the sum of the positron mass and the neutron proton mass difference. Applying the approximation E e . « En χ mnc2 valid for the low energies of reactor neutrinos one finds for the total (vcp) cross section [50]: ^P(£e +)=

X £e +

~

X \M\*

(10)

Μ denotes the transition matrix element of the (vep) reaction and is evaluated in analogy to the matrix element of the decay of free neutrons: • -

2

*V f>,»>y

,1!)

This formulation allows a value for \Ji\2 to be derived from the experimental value of the neutron mean life tn, with the present world average value [51] of tn = 889.1 + 2 . 1 sec. The phase space factor/in eq. (11) is derived analytically and the quoted value of / = 1.71465 + 0.00015 includes Coulomb-, finite mass and radiative corrections and effects from a finite charge radius of the nucleus [52].

284

Jean-Luc Vuilleumier and Gabriele Zacek

Inserting the individual constants into eq. (10) the (v e p) cross section can be written as: σ ν« Ρ ( £

+j=

(9

43 + 0.17) χ (£ e , - m

2 t

) +

mec2)2 - m2c4 χ 1 0 " 4 4 c m 2

F o r a more accurate evaluation several corrections at the percent level have to be made. These corrections include contributions from weak magnetism which is due to the difference in the magnetic moments of neutron and proton and radiative corrections of order a not yet included in the calculation of the phase space factor of neutron decay [53,54],

2. Neutrino capture on heavy nuclei The electron neutrino capture reaction on heavy nuclei, ve + (Ζ, N) ->• e~ + (Z + 1, TV — 1), is one of the methods suitable for the detection of solar neutrinos. A neutrino capture reaction is tagged by the presence of the daughter nucleus ( Z + \ , N — 1), which is identified through its characteristic decay modes. The cross section for the absorption of electron neutrinos is generallized from eq. (10) and reads: j i J U

'

( £ t ) =

i

x

£

t

Λ / £ 2 — m 2 c 4 χ F(Z,Ee

) χ \Ji\2

(12)

where £ c is the energy of the emitted electron, \ Ji\2 denotes the transition matrix element and F(Z, Ee ) refers to the integral Fermi function which includes Coulomb and nuclear size corrections. Like in the case of neutron decay the transition matrix element Ji can be expressed through the ft-value of the respective neutrino capture reaction. However, in calculating the phase space factor / the more complex nuclear structure has to be taken into account. To avoid this complication ft-values are extracted from experimental data, which measure either the inverse electron capture process or employ low energy proton beams to measure the cross section of the corresponding (p, n) reaction. Of course in both cases corrections have to be applied to convert the measured cross section values into the corresponding neutrino absorption reaction.

B. Neutrino deuteron scattering Neutral and charged current scattering modes are possible for the interaction of neutrinos with deuteron atoms [55], While the N C reaction (vd -> pn v) is not sensitive to the incident neutrino flavour the C C reaction (v c d->ppe~ resp. v e d - > n n e + ) only occurs for electron neutrinos. In case of a neutral current interaction the reaction threshold equals the binding energy δ = 2.226 MeV of the deuteron. F o r charge currents the threshold energy differs for incident neutrinos ( E l h r = 1 . 4 4 MeV) or antineutrinos (£ thr = 4.03 MeV) and reflects either the transformation of a neutron into a proton or the transformation of a proton into a neutron.

Neutrino Detectors

285

Folding the energy spectrum of reactor neutrinos with the (vd) cross sections yields per fission a value of a(CC)= 1.2 χ 10" 4 4 cm 2 for charged current scattering and a(NC) = 2.9 χ 10" 4 4 cm 2 for neutral current scattering [16]. Due to threshold effects and a smaller overlap of initial and final state nuclear wave functions, the (v e d) cross sections are lower than the (vep) cross section by about a factor 20. In solar neutrino experiments where the high threshold essentially only allows the detection of 8 B neutrinos (v e d) cross sections lie in the region of 10" 4 2 cm 2 . For the study of neutrino oscillations the (vd) reaction provides the interesting aspect that the flavor independent neutral current mode can serve as monitor reaction for the total incident ν flux, while possible oscillation effects will turn up in the flavor dependent charged current channel. This feature was exploited by one reactor neutrino experiment [56] and will constitute an important ingredient of the solar neutrino experiment SNO [57] in future.

C. Neutrino electron scattering The elastic and purely leptonic scattering of neutrinos on electrons, ve->ve has developed in recent years to become an important tool for neutrino detection. In the case of solar neutrinos neutrino electron scattering allows in contrast to the radiochemical experiments with Gallium and Chlorine the identification of individual events. This is extremely interesting for the study of time dependent cosmic neutrino sources (like e.g. SN87A) since the observation time of the events can be correlated to the occurrence of the neutrino burst. The kinematics of neutrino electron scattering corresponds to the simple case of the scattering of a massless particle, the neutrino, on a particle with mass me and the relation between the electron scattering angle 0 e , the neutrino energy Ev and the kinetic energy Te of the scattered electron in the laboratory frame of reference is given by:

(13) For high neutrino energies ( E v , T e » m e ) relation (13) reduces to Θ 2 ^(2m e /T e ) χ (1 — TJEV) and electron are preferably scattered into the forward direction. Since the signature of a (ve) scattering events is just the appearance of a single electron, methods have to be developed to recognize this electron against prevailing background sources. One possibility is to correlate the direction of the recoil electron to the direction of the incident neutrino. This method is adopted by the Kamiokande collaboration where neutrinos of solar origin are required to be forward scattered around the sun/earth axis [19]. While only electron neutrinos resp. electron antineutrinos can be detected in experiments based on inverse ß-decay modes, neutrinos of all flavours react via neutrino electron scattering processes. The cross-section is, however, flavour dependent. This is because vc — e~ scattering has contributions from neutral and charged currents, while νμ — e~ and ντ — e~ scattering are purely neutral current processes. The cross sections

286

Jean-Luc Vuilleumier and Gabriele Zacek

can be derived in a straight forward way within the framework of the standard model of electroweak interactions [58] and depend on one free parameter, the weak mixing angle sin 2 6>w. Assuming a value of 0.23 for the weak mixing angle the total (vee) cross section a m o u n t s to: ata\vee)/Ev

cm 2 = 9.49 χ Ι Ο " 4 2 — — GeV

ff,0,(vee)/£v

cm 2 = 4.27 χ Ι Ο " 4 2 — GeV

For muon or tau neutrinos the cross section is reduced by about a factor six with respect to ff(vee).

IV. Neutrino detectors at nuclear reactors A. Detectors based on the v e p - » e + n reaction To exploit the full potential of the v e p - > e + n reaction in neutrino experiments at nuclear reactors, detectors are generally designed to record both reaction products. While liquid scintillation counters directly detect the emitted positrons, neutrons first have to be thermalized before they can be detected efficiently by means of neutron capture reactions for which a variety of nuclei are available. If possible the event is also validated by requiring a time and position correlation between positron and neutron detection. D u e to thermalization and diffusion the signal from the detection of the neutron will arrive later than the prompt positron signal. An important design feature is the capability of background reduction. In most cases reactor correlated background can be avoided by adequate shielding materials. Background events from high energetic particles in the cosmic radiation are more serious and have to be reduced with the help of active and passive shielding at least to the level of the expected neutrino event rate. Cosmic and also accidental background can be measured during reactor-off periods.

1. The first neutrino

experiments

In 1953 F. Reines and C. L. Cowan reported for the first time on the "detection of free neutrinos" at the nuclear reactor in Hanford (USA) [59]. One single 10ft 3 large cylindrical tank filled with cadmium loaded liquid scintillator was used as a neutrino detector. The target volume was viewed by photo-multipliers mounted uniformly around the cylindrical walls and signature for a neutrino event was the occurrence of one scintillator pulse generated by the positron followed by a delayed second pulse due to the γ-cascade initiated by neutron capture on cadmium atoms. The final result of this first measurement had only a significance of about one standard deviation and was quoted as a "probable" detection of free neutrinos. To improve on the event identification and thereby to obtain a more reliable result a second experiment was designed and installed at the Savannah River plant of the U.S. Atomic Energy Commission [60], The detector consisted of an array of three scintilla-

Neutrino Detectors

287

tion counters interspaced with two target tanks containing a water solution of cadmium chloride which provided the protons for the neutrino interaction and the cadmium atoms necessary for neutron detection. With this the background event rate could be suppressed by requiring a spatial correlation between positron and neutron detection. This second experiment was in operation for 1371 hours including the reactor down time in which the accidental background could be measured. The evidence for the neutrino signal was now much stronger: 2.88 + 0.22 counts per hour were attributed to neutrino interaction and the signal to background ratio was stated as 3 to 1. In a third experiment [61] F. Reines and collaborators practically went back to their original design of the detector installed at Hanford and used one tank of cadmium loaded liquid scintillator. In comparison to the Hanford detector this detector was superior since it contained a factor 6.5 more target protons (8.3 χ 10 28 ) and several technical improvements including a more elaborate shielding. After background subtraction a reactor correlated signal of 36 + 4 events per hour was seen. Based on this event rate a value of σ = (11 + 2.6) χ 10~ 4 4 cm 2 was quoted for the total (v e p) cross section, averaged over the reactor spectrum.

2. A neutrino detector based on a Gadolinium loaded

scintillator

In a 2nd generation of neutrino detectors Nezrick and Reines at Savannah River again adopted a sandwich type detector design [62]. O n e central tank (27 cm φ, height 7 cm) which was flanked on both sides with N a l counters (29 cm φ, height 7.6 cm) contained the neutrino target: a proton rich liquid scintillator which was loaded with gadolinium for neutron detection. Signature for positron detection was a signal above 0.5 MeV from the central scintillation counter in prompt coincidence with pulses from both N a l . The pulses from the N a l counters were expected to be caused by the two p h o t o n s following positron annihilation and had to correspond to an energy of 0.511 MeV. T o identify the corresponding neutron interaction it was required that the delayed coincidence signal followed the first pulse within 2.8 μ sec to 50 μ sec to cover the timespan necessary for neutron thermalization and the deexitation of the gadolinium nucleus after neutron capture ( « 20 μ sec). The delayed signals in the N a l counters had to be above 0.75 MeV to guarantee the detection of the Gd(n,y) Gd capture in which, on average, four 2 MeV γ-rays are generated. The accidental background was monitored in a time window of 50 μ sec to 100 μ sec following the positron signal. Pulses from the photomultipliers were processed in a logic electronic circuit and a signal sequence showing the correct delayed coincidence structure triggered an oscilloscope and a camera control unit. The oscilloscope showed the p r o m p t signals from the central target detector and the two N a l counters together with the delayed pulses from all three counters. During reactor operation 28 pictures were taken per hour which subsequently were scanned and analyzed. Less than 1 % of these pictures showed the characteristic delayed coincidence sequence. The overall detection efficiency for a correlated positron/neutron pair ranged from 3.3% to 2.3% for positron energies of 0.75 MeV to 8.25 MeV respectively. The largest losses resulted from positrons escaping the target region or interacting in the lucite

288

Jean-Luc Vuilleumier and Gabriele Zacek

walls of the target tank. Altogether the experiment took data during 2484 hours of reactor operation and during 357 hours with the reactor being shut down. After background subtraction a neutrino signal rate of 0.187 ± 0.021 events per hour was measured which yielded a value of (0.94 + 0.13) χ 10 4 3 cm 2 for the total (v e p) reaction cross section.

3. The experiment at the Gösgen nuclear power plant A different variant of a detector system consisting of an alternating array of liquid scintillation counters and 3 H e multiwire proportional chambers was developed for the neutrino oscillation experiment at the 2800 M W t h nuclear power station in Gösgen (Switzerland) [5]. The liquid scintillator counters provided the target protons for the neutrino interaction and at the same time worked as positron detectors. Neutrons first thermalized in the scintillator oil and then diffused into one of the adjacent 3 He counters where they were detected with very high efficiency. In both detector systems time and position of positron and neutron detection were recorded. For a valid neutrino interaction both events had to be registered within 24 cm and within a time interval of 250 μβ. These selection criteria were based on the thermalization and diffusion times of the neutron and the length of the path the neutron traversed before reaching one of the adjacent neutron counters. In two other experiments (ILL in Grenoble [63], Bugey 1,11 [64,65]) similar detectors were employed which, however, had inferior event identification possibilities. Figure 3 shows the set-up of the Gösgen neutrino experiment. The 3771 of organic liquid scintillator provided (2.41 + 0.02) χ 10 28 target protons and was contained in 30 transparent lucite cells (88 χ 20 χ 9 cm 3 ). Six cells were stacked on top of each other and made up an array of five target cell planes. For positron detection these scintillator cells were equipped at each end with two photo-multipliers. The length of the target cells was chosen so as to obtain a good light collection efficiency and the target cell width was a compromise between the volume available for neutron thermalization and unwanted capture of neutrons in the scintillator oil. For a valid neutrino event the light pulse in the target cells is caused by the ionization energy loss of a positron. However, a fast neutron created e.g. by a cosmic ray muon in the material surrounding the detector could hit a proton in the scintillator, thermalize and diffuse into one 3 H e counter, while the proton would generate a light pulse in the target cell. This process could simulate good neutrino event but was efficiently suppressed by using pulse shape discrimination techniques on the shape of the scintillation light signal. The scintillation light pulse has a fast (4 ns) and a slow (100ns) component. The fraction of light which appears in the fast component essentially depends on the ionization density of the traversing particle. Since the ionization density of a positron is smaller than that of a proton it is possible to discriminate between the shapes of the two light pulses. Selecting positron events according to the shape of the detected light pulse allowed a reduction of the background by a factors 4 to 5. The intrinsic energy resolution of the scintillation counters was measured using forward scattered C o m p t o n electrons of various calibration sources which were immersed into the liquid scintillator. However, in the case of neutrino interactions

Neutrino Detectors

289

several additional effects deteriorated this experimentally determined energy resolution: escape of positrons f r o m the sensitive target volume, the additional detection of bremsstrahlungs p h o t o n s or of the two 511 keV annihilation p h o t o n s a n d c o n t r i b u tions from neutrino interactions in the lucite walls of the target cells. All these effects were taken into a c c o u n t when modelling the energy response of the scintillation counters in the Gösgen experiment [66]. T h e achieved energy resolution was a b o u t 2 2 % at a positron energy of 3 MeV. The event positron along the horizontal dimension of a target cell was d e t e r m i n e d from the time difference of the signals recorded by the photo-multipliers at b o t h ends of the cell and a position resolution of a b o u t 8 cm was obtained. In the vertical direction the event was localized by the discrete a r r a n g e m e n t of the scintillation counters. N e u t r o n s emerging f o r m the v e p - » e + n reaction with an energy of several keV thermalized within 10 μβ in the liquid scintillator a n d diffused into one of the a d j a c e n t multi wire p r o p o r t i o n a l c h a m b e r s filled with 3 H e gas. Altogether four 3 H e c o u n t e r s with dimensions of 120 χ 8 χ 90 c m 3 were m o u n t e d in between the five target cell planes. T h e c a p t u r e process of thermal n e u t r o n s by 3 H e has the large section of 5330 barns and yields a p r o t o n a n d a triton with a reaction Q-value of 765 keV: η + 3 H e - » ρ + 1 + 765 keV. T h e 3 H e wire c h a m b e r s were operated in the p r o p o r t i o n a l region at a t m o s p h e r i c pressure with a mixture of 95 vol % 3 H e a n d 5 vol % C 0 2 . A central wire plane represented the a n o d e while highly polished a n d very thin (neutron capture!) c h a m b e r walls are g r o u n d e d and served as cathodes of the detector. T h e position sensitivity along the wires of the 3 H e c o u n t e r s was accomplished by charge division and a position resolution of a b o u t 16% 8 cm) was obtained which matched the extension of a p r o t o n / t r i t o n event in the chambers. T h e energy s p e c t r u m recorded in the 3 H e counters showed a clear peak at 765 keV c o r r e s p o n d i n g to the detection of a fully contained p r o t o n / t r i t o n pair with an energy resolution between 15% and 2 0 % for the four chambers. The detection efficiency for slow neutrons in the c o u n t e r s (ε = 0.217 + 0.008) is an i m p o r t a n t quantity a n d was measured with a S b - B e source. C o m b i n i n g this value with the efficiencies of positron detection and of the selection criteria a final a n t i n e u t r i n o detection efficiency of a b o u t 16% was reached. Since the n e u t r o n counters served as trigger counters in the neutrino experiment it was of the highest i m p o r t a n c e in the construction of the 3 H e c h a m b e r s to keep t h e m free of any n a t u r a l radioactivity. All materials employed were selected for low α activity. Finally a b a c k g r o u n d rate due to n a t u r a l radioactivity of only 1.2 cts/min in the energy window of interest was reached. F o r the scintillation counters the p h o t o multipliers were pre-selected with respect to 4 0 K c o n t a m i n a t i o n (1.46 M e V γ) of the PM-glass in order to reduce the accidental b a c k g r o u n d . Differential shielding tests h a d shown that there was n o b a c k g r o u n d associated with the operation of the reactor and thus the only remaining external b a c k g r o u n d source was due to cosmic radiation. The aim of the detector shielding was to suppress the cosmic induced fast n e u t r o n count rate a n d secondary b a c k g r o u n d sources like 4 0 K activity f r o m the shielding itself to the level of the single b a c k g r o u n d rates of the individual counter systems (single rate in the target cells « 300 cts/sec; single rate in one 3 H e c h a m b e r ä 0.3 cts/sec). F o r this p u r p o s e the central p a r t of the neutrino detector

290

Jean-Luc Vuilleumier and Gabriele Zacek

Fig. 3. Set-up of the Gösgen neutrino detector system. The central neutrino detector which is located during operation inside the concrete house (H) is shown in its rolled out position. (1) central detector unit; (2) active veto system; (3) rails to move the detector; (4) steel door; (5) water tanks; (6) movable concrete door.

was surrounded from outside to the inside by the following shielding materials: 2 m of concrete + additional 2 m of concrete above the detector; 15 cm of iron; 20 cm of water contained in steel tanks; 5 mm of B 4 C plates (see Fig. 3). The concrete shield was sufficient to reduce the nuclear component of the cosmic radiation by a factor 10 4 , however, 4 0 K present in concrete introduced a large γ activity of about 10 5 cts/sec. The iron layer following the concrete shield had to absorb this background and its thickness of 15 cm was a compromise between background reduction and the requirement of the least amount of material for additional neutron production by traversing high energetic particles. Nevertheless water and B 4 C had to be added to moderate and absorb secondary neutrons generated in the iron shield. The highly penetrating muonic component of the cosmic radiation cannot sufficiently be reduced by passive shielding materials. Thus an active veto system of 12 cm thick scintillation counters hermetically surrounded the neutrino detector and was working in anti-coincidence to the central detector unit. The m u o n detection efficiency was only limited by the mecahnical tightness of the construction and was better than 99.8%. Since neutrons produced by muons could thermalize in the structure material and diffuse into one of the 3 H e counters, a 320 μβ long veto gate was generated by each muon which traversed a veto counter and events registered in the 3 H e chamber within this timespan were rejected. Furthermore, to recognize bremsstrahlung of secondary electrons, a second 10 μβ long gate was applied in anti-coincidence to signals from the target cells.

Neutrino Detectors

291

In the Gösgen experiment data were taken at three different distances (37.9m, 45.9m, 64.7m) from the reactor core and about 10 4 v c were collected in each measuring position. The data, absolute event rate and positron spectral distribution, were consistent with the absence of neutrino oscillations and the limits: Am2 < 0 . 0 1 9 (90% c.l.) for maximum mixing and sin 2 2Θ ν < 0.21 (90% c.l.) for A m 2 > 5 e V 2 were quoted. The excluded area in the Am2 — sin 2 2 0 v plane is shown in Fig. 2.

4. The neutrino experiment at the reactor in Rovna the Kurtchatov group

(Ukraine),

In the first experiment at Rovna one tank containing 240 1 of liquid scintillator loaded with gadolinium made up the central part of the neutrino detector and was installed at a distance of 18.06 m from the reactor core [67,68]. Neutrino interactions were identified by the delayed coincidence between the positron signal and the γ-rays emitted during deexitation of the gadolinium nucleus. The usage of only one detector tank reduced the inefficiencies due to positron escapes and neutrinos interacting in the detector walls and an overall efficiency of 31.1 + 1.8% was reached. About 15 χ 10 3 v e interactions were recorded during a running time comprising 57.2 days of reactor operation and 20.1 days of background data taking. A quite particular detector concept was adopted for the second Rovna experiment [69,70], In contrast to all other approaches where the signature for a neutrino interaction was based on a delayed coincidences between positron and neutron detection here only the neutrons produced in the (v e p) reaction were detected. Although the possibility of measuring the neutrino spectrum is lost, this approach has the potential of reaching a high neutrino detection efficiency and thus a significant increase in statistics. Neutrino target and neutron moderation material were 136.4 kg of polyethylene subdivided into 11 blocks. Into each of these blocks 12 3 H e filled proportional counters (3.2 cm φ, 1 m length) working at a pressure of 4 atm were inserted. The mean lifetime of neutrons in the detector was about 50 μ sec. T o reduce background due to the intrinsic activity of the 3 H e counters a tight energy window was set around the characteristic 765 keV peak which introduced an inefficiency of 30%. The entire neutron detection efficiency was determined to be about 52% where the largest losses ( ~ 2 0 % ) were attributed to absorption of neutrons in the moderator and in materials surrounding the central detector. Though the overall event detection efficiency significantly exceeds those of all other experiments the major problem in the "pure neutron" approach resides in the background suppression and a careful selection of all detector materials was necessary. To reduce the cosmic ray background the whole detector was hermetically surrounded by borated polyethylene and an active scintillation counter was mounted on top of the detector and was operated in anti-coincidence with the 3 He-counters. In addition the whole detector set-up was installed in a specially built low background laboratory. For the analysis of the first data taken up until summer 1983 only signals from a central part of the detector containing 36% of the total available protons were used. For the averaged cross-section of the (v e p) reaction, normalized per fission, the group quotes a result of (5.79 + 8 % (stat.) + 7 % (syst.)) χ 10~ 4 3 cm 2 /fission.

292

Jean-Luc Vuilleumier and Gabriele Zacek

A very similar detector concept was adopted for the 'Integrating Neutrino Detector (IND)'by a group of Kurtchatov Institute [71], The central detector part (80 χ 80 χ 97 cm 3 ) was constructed from plexiglass which provided 2.78 χ 10 28 target protons and also serves as neutron moderator. 105 3 H e filled proportional counters were inserted into the plexiglass cube and an antineutrino detection efficiency of 31.1 ± 0 . 9 % was obtained. The site of the I N D allowed the antineutrino fluxes to be monitored from two reactors located at distances of 32.8 m and 92.3 m respectively. It was operated during a period of 1200 days and a total number of about 10 5 (v e p) events were detected. The averaged reaction cross section is quoted as: (6.19 ± 0.20(stat.) + 0.30 (syst.)) χ 10~ 4 3 cm 2 /fission. Results from a third "pure neutron" approach were reported in September 1990 by the Kurtchatov group [72]. The neutrino detector, a hexagonal prism made out of aluminium ( ~ 50 cm φ, 110 cm length), filled with 460 kg of granular polyethylene, provided 4.1 χ 10 28 target protons and was equipped with 90 3 H e proportional counters. It was sensitive to the neutrino flux from three reactors which were located at distances of 57 m, 57.6 m and 231.4 m from the low background neutrino laboratory. Limits on oscillation parameters were derived, from a comparison of the spectra at different distances, and are shown in Fig. 2. Combining the data from its two experiments, the Kurtchatov group quotes an average cross section for the (vp) reaction of (6.11 + 2 . 5 % (detector) + 1.8%(reactor)) χ 10~ 4 3 cm 2 /fission, in agreement with expectations.

5. The 3rd neutrino experiment at the Bugey power plant

(France)

For the set-up of Bugey III [73,74], performed at one of the four 2800 M W t h reactors of the Bugey plant, a new design was adopted based on a liquid scintillator loaded for neutron detection with a small amount (0.15% in weight) of highly purified 6 Li. The liquid scintillator was contained in a 6001 tank which was optically segmented into 98 cells (8.6 χ 8.6 χ 8.7 cm 3 ) which were equipped at both ends with photo-multipliers. D a t a is taken with three such detector modules. O n e of them is installed in a position close to the reactor core (16 m) while the two others are located 40 m away from the v e source. This set-up allows a simultaneous measurement of the reactor neutrino spectrum at two distances with identical detector modules and thus uncertainties due to the reactor operation can be reduced. The prompt signal from positron detection is confined to one single detector cell and the delayed neutron signal is expected to appear within ~ 29 μ sec in one of the neighbouring five cells. Thermal neutron capture by 6 Li atoms produces a α-triton pair with a total energy of 4.8 MeV: η + 6 Li->oe + 1 + 4.8MeV. However due to the quenching of light in the scintillator the effective light output only corresponds to that of a 530 keV electron. Pulse shape discrimination techniques are therefore necessary to distinguish positron and thermal neutron capture signals. F o r further event identification the position of the event in the longitudinal direction is reconstructed by analyzing the arrival time and the amplitudes of the scintillator pulses at both cell end faces. Adopting values of 9 0 % and 70% for the positron resp. neutron detection efficiency

Neutrino Detectors

293

a neutrino rate of 90 events per hour is expected for the 16 m position and the two detector modules at 40 m should detect about 28 (v e p) events in one hour.

B. Detectors based on the vd capture reaction The U C Irvine group built a detector for the study of v e d scattering and installed in at the 2000 Μ W t h Savannah River reactor, at 11.2m from the core [56]. The detector consists of a vessel filled with 268 kg of D 2 0 . Ten cylinderical proportional counters, filled with 3 He, are immersed in the D 2 0 . They serve to detect the neutrons produced in the neutral current (NC) reaction ve + d - > v e + n + p and in its charged current (CC) counterpart v e + d - > e + + η + n. Both types of reactions can thus be seen, although the background is much less for the C C reactions, since a coincidence between two neutrons is required. To reduce the background, the detector is protected by several shielding layers: 10.2 cm of Pb, 0.1 cm, of Cd. All this is surrounded by a 30 cm thick liquid scintillator serving as veto counter. The efficiencies were determined by Monte Carlo simulation. The efficiency to detect two neutrons was found to be 0.112 + 0.009, and for one neutron 0.32 + .02. D a t a were accumulated for 95 days reactor on, and 66 days reactor off. The difference gives the reactor associated signal, about 70 per day for single neutron events (signal to background ratio of 1:6), and 4 per day for two neutron events (1:13) [75]. Taking into account the detection efficiencies, the ratio of C C events to N C events was computed: (CC/NC) m c a s u r c d = 0.282 ±0.089 It essentially agrees with expectations [76], This can be seen best by looking at the ratio of measured to calculated ratio, which does not depend on the overall normalization of the reactor spectrum. It is close to 1: (CC/NC) m e a s u r c d

=

(CC/NC) c a l c u l a t c d These reactions have also been studied by the Kurtchatov group [77], They built a detector consisting of a 406 1 tank of heavy water surrounded by graphite blocks, acting as the neutron moderator and reflector. The outside dimensions are 1.8 χ 1.8 χ 1.5 m. A total of 151 holes are made through the graphite, including 91 also going through the tank. Of these holes, 150 are instrumented with low background proportional counters for neutron detection. The last hole is used for calibration. The detector is further surrounded by various shielding layers: 8 cm of B 4 C, 24 cm of boron loaded polyethylene, and, on the sides, 20 cm of steel. On top, there is a plastic

294

Jean-Luc Vuilleumier and Gabriele Zacek

scintillator, to veto the cosmics. The neutron mean lifetime in the detector is τ = 190 + 2 μβ. The efficiency for detecting one neutron is 75.8 + 2.3% in a time window of 4τ, that for detecting two neutrons is 59.6 + 1.6%. D a t a were taken between two reactors, at distances of 34 m and 87.7 m. The far reactor was always on, while the near reactor was off part of the time, periods during which the background was measured. Part of the boron carbide shielding, which was found to create more background that it was suppressing, was removed during the course of the experiment. F o r one neutron events, the signal to background ratio was found to be around 1:40. In spite of this, a reactor correlated effect was seen, and an average cross-section has been reported for the neutral current reaction: ^

= (3.0+ 1.0)-10" 4 4 cm 2 /fission.

For two neutron events, the signal to background was 1:6, and the cross-section for the charged current cross section was determined to be: ffcc = (1.1 ± 0.3)· 1 0 " 4 4 cm 2 /fission. Both cross-sections are in good agreement with expectations [76], the ratio of ratios being: (CC/NC) m e a s u r e d

= 0 92 + o 40

(CC/NC) c a l c u l a l c d

C. Detectors for v e e

scattering

Although of fundamental importance, the scattering v e e" has been unambiguously observed in only two reactor experiments, so far. They are presented in the following.

1. The experiment at the Savannah River reactor The U C Irvine group led by F. Reines [78] built the first detector dedicated to the study of elastic v e e ~ scattering. The detector was operated successfully at the Savannah River Plant (SRP) in the mid-seventies, where the reaction was observed for the first time. The aim was, in those days, to verify the predictions of the standard model. The detector consisted of a 15.9 kg plastic scintillator, coarsely segmented and surrounded by a N a l counter, a P b shield and a liquid scintillator to veto the cosmics. The signature for a good event was given by a single count in one of the elements of the plastic scintillator with nothing in coincidence. The N a l gave a good anti-Compton efficiency allowing an efficient suppression of the γ background. Events from the ve + p - > e + + n reaction in the plastic (200 events · day ~ M) were efficiently identified by the detection of the annihilation γ rays a n d / o r by the delayed neutron capture signal in the plastic scintillator or the N a l , and rejected.

N e u t r i n o Detectors Table 1. Event rate in the Savannah River v e e T(MeV)

Events/day

1.5-3 3-4.5

Reactor on 45.1 + 1.0 2.4 + 0.19

295

experiment

Reactor off 39.2 + 0.9 1.2 + 0.14

on-off 5.9+1.4 1.2 + 0.25

The detector was placed at 11.2 m from the core of the reactor operated, at that time, at a power of 1800 M W t h , so that the neutrino flux was around 1.9· 10 1 3 c m ~ 2 s _ 1 . Since the cross section is small and the signature rather poor it was necessary to reduce the background from natural activities to a minimum by selecting clean materials for all the components. Events were recorded during 64.6 days reactor on and 60.7 days reactor off. The count rates in two bins of electron recoil energy Τ are given in Table 1. In spite of the low singles background the signal to noise ratio is somewhat marginal. These data were compared with standard model predictions based on eq. 7 and on the then accepted reactor spectrum, resulting from a calculation [79]. Agreement was found and the value sin 2 0W = 0.29 + 0.05 was derived. This is consistent with the best value to date obtained by the C H A R M II collaboration [80] in ν μ ε~ scattering, another purely leptonic process: sin 2 0W = 0.239 + 0.011. We now know however that the reactor spectrum used in this original analysis was much too hard. Vogel and Engel [33], using the best determination at present of the reactor spectrum and fixing sin 2 0W to the presently accepted value found that the measured rates in the two energy bins given above are 1.35 + 0.4 and 2.0 + 0.5 times larger than the expected ones. Taken literally this discrepancy indicates new physics beyond the standard model. O n e can reconcile the data with eq. 7 by introducing a neutrino magnetic moment μ ν = (2-4)· 1 0 " 1 0 , of the right order for the spin flip mechanism in the sun to occur! Of course, the experiment is very difficult, and the discrepancy may be due to an instrumental problem. It therefore, appears important to perform new experiments to clarify the situation.

2. The Kurtchatov

experiment

More recently, a group from the Kurtchatov Institute in Moscow has also successfully observed ve~ scattering [81]. The detector consists of seven identical cells, and filled with a C 6 F 6 based liquid scintillator (103 kg in total), with low hydrogen content, which serves as active target material. These cells are viewed by two photo-multipliers, one on each side, connected by long light guides to suppress the background from the glass. All the materials chosen are radiochemically very clean, to reduce the background from natural 4 0 K , 2 3 2 T h and 2 3 8 U activities, and are essentially hydrogen free, so that there is little background from the reaction ve + p - > e + + n. The detector is surrounded by various shielding layers to reduce the background from local activities. A plastic

296

Jean-Luc Vuilleumier and Gabriele Zacek

Table 2. Event rate in the Kurtchatov t>ee T(MeV)

Events/day

3.15-5.17

Reactor on 8.27 + 0.18

experiment

Reactor off 7.49 + 0.31

on-off 0.78 + 0.36

scintillator, placed on t o p of the shieldings, allows the b a c k g r o u n d from the cosmics to be reduced. T h e neutrino flux, at the detector site, is 3.4· 10 1 2 c m ~ 2 s _ 1 . So far, d a t a have been taken for 250 days with reactor on, a n d 80 days with reactor off. T h e difference gives the reactor associated signal, which is clearly present above 3.15 MeV electron recoil energy, as can be seen f r o m Table 2. It is estimated that 0.1 event/day comes from the b a c k g r o u n d of the reaction v e + p - > e + + n, the remaining 0.68 + 0 . 3 6 being due to v e e~ scattering. This rate is compatible with expectations, obtained with sin 2 0w = 0.23, and leads to the limit μ ν < 2.4· 1 0 ~ 1 0 for the neutrino magnetic m o m e n t . This result is consistent with that f r o m S a v a n n a h River. T o improve, it a p p e a r s necessary to reduce the b a c k g r o u n d at low energy in order to lower the threshold on the electron kinetic energy.

V. Solar neutrino detectors It is well k n o w n that experiments to detect solar neutrinos are very difficult. Large masses are necessary, because of the low cross-sections and fluxes. T h e energies are low, of order of a few MeV, and, as a consequence, the b a c k g r o u n d problems from natural activities a n d f r o m cosmogenic activations, are tremendous. So far, four experiments have successfully observed solar neutrinos. Three are radiochemical experiments ( 37 C1, G A L L E X , SAGE), in which one looks in a large mass of target material, after a given exposure time, for nuclei activated by n e u t r i n o interactions. The fourth experiment ( K a m i o k a n d e ) directly detects solar neutrinos. These experiments are described in the b o o k by J. Bahcall [6], A brief s u m m a r y is given here.

A. The radiochemical method 1. The

37

CI

experiment

The 3 7 C1 experiment [18,35] was the first to see solar neutrinos. It has been in operation since 1970, with few interruptions. It uses the detection reaction ve +

37

Cl->e~ +

37

Ar

(£ l h = 0.814 MeV).

The threshold energy £ t h is low and, in principle, the experiment is sensitive to neutrinos f r o m all solar reactions, except pp. However, high energy neutrinos f r o m 8 B decay (Q = 14.6 MeV), which are expected to contribute 7 7 % of the total event rate,

Neutrino Detectors

297

should dominate. Chlorine is a good target material since it can be produced in large quantities, and since the isotopic abundance of 37 C1 is 24.2%. The 3 7 Ar nuclei decay back to 37 C1 by electron capture: 37

Ar + e~ -> 37 C1 + v e ,

T 1/2 = 35d.

With 90% probability, a Κ electron is captured. An Auger electron is then emitted, with an energy Ε = 2.8 keV, sufficiently large to make detection possible. The detector consists of a tank filled with 6151 of liquid perchlorethylene, C 2 C1 4 (2.2 χ 10 30 atoms of 37C1). It is located in the Homestake mine at a depth of 1400 m of rocks, or 4000 mwe (meter water equivalent). This largely reduces the background from cosmic activation. At the beginning of each exposure, about 0.2 cm 3 (STP) of isotopically pure 3 6 A r or 38 Ar carrier is dissolved in the C 2 C1 4 . The tank is then exposed for a period of typically one to three months. After this, 4 χ 10 5 1 of He are circulated, first bubbling through the C 2 C1 4 to extract the Ar, then through various filters, and finally through a cold trap at liquid nitrogen temperature with active charcoal, where the argon is captured. It was found that 9 5 % of the 3 6 , 3 8 A r carrier is extracted. It is assumed that 3 7 A r atoms induced by solar neutrinos are extracted with the same efficiency. The recovered gas, to which 5 to 10% methane is added, is introduced into a small proportional counter (0.3-0.5 cm 3 ). O n e then counts the 3 7 Ar atoms by looking at the 2.8 keV Auger electrons, for a period extending over several months. The energy resolution is about 25% F W H M . Event candidates must not only have the right energy, but also a fast rise time, corresponding to a short electron track. The detection efficiency is a r o u n d 0.57 (0.85 geometrical acceptance times 0.67 acceptance for the energy and rise time window). The event rate, in function of counting time, is fitted varying a flat background with, on top, an exponential spectrum with the decay time o f 3 7 Ar. This allows a good reproduction of the measured spectra, and determines the background and signal rates. Several detectors, with better and better background, have been used over the years. The detectors are now operated in the underground lab itself. Presently, the background rate, corresponding to the event rate at the end of the counting period, is around 0.01 count per day only, but is was around 0.02 count per day on average for the period from 1970 to 1992. In the beginning of the counting period, the event rate is roughly 0.16 per day. Integrating over the entire counting period, and subtracting the background, about 7 3 7 Ar atoms are detected. Normalized to the exposure time, this corresponds to a detection rate of order 0.1 per day. Taking into account the acceptance (0.95 for the extraction, 0.57 for the detector acceptance), and correcting for the number of 3 7 Ar which decay before counting begins, about half, as well as for the Κ capture probability (90%), one gets a n 3 7 Ar production rate for all runs until 1992 of 0.50 per day. In total, some 600 3 7 Ar atoms have been observed. From this, the number of 3 7 Ar nuclei produced by cosmic activation must be subtracted. It was determined by exposing a 600 gallon tank of C 2 C1 4 at shallower depths, and extrapolating to the depth of the lab. A theoretical estimate, normalized to the measured number of muons in the lab, gave a similar value. The adopted rate is 0 . 0 8 + 0 . 0 3 count per day. With this, the total production rate of 3 7 Ar by solar neutrinos becomes 0.42 + 0.005 count per day. Expressed in solar neutrino units (SNU,

298

Jean-Luc Vuilleumier and Gabriele Zacek

interaction per 10 3 6 target nuclei per second) it is (2.3+0.3) SNU. As stated in a preceding section, this is a factor two to four less than the rates predicted by calculations based on the standard solar model: 8.0 + 1.0 S N U [7], 6.4 + 1.4 S N U [8], The origin of this discrepancy is still a matter of speculation. It should be added that a time variation of the signal, anti-correlated with the solar activity measured from the number of sunspots, is suggested by the data. The count rate was low around 1979, when the solar activity was high.

2. The

71

Ga detectors: GALLEX

and SAGE

The G A L L E X experiment [22] uses 7 1 G a , another good candidate, as target material and is conceptually similar. The detection reaction is ve +

71

Ga->e" +

71

Ge

(£ t h = 0.233 MeV).

The low threshold makes the detection of p p neutrinos possible. In fact, pp and pep neutrinos are expected to contribute 56% of the total rate. The flux from these neutrinos is much less uncertain than that of higher energy neutrinos. After production, 71 Ge nuclei decay to 7 1 G a by electron capture (T 1 / 2 = 11.4 d). The natural abundance of 7 1 G e is 39.7%. The experiment is being performed in the G r a n Sasso underground lab, at a depth of 3500mwe. The target consists of 30.31 of gallium in the form G a C l 3 in an 8.13 Μ aqueous solution acidified to 2 Μ in HCl, contained in a single tank. The concentration is such that the Ge atoms will be in the form of GeCl 4 , which is volatile. It can thus be separated by bubbling an inert gas. In GALLEX, this is done efficiently by purging 1900 m 3 of nitrogen during 20 h. After an extraction, the next exposure starts. First, a small a m o u n t of germanium (about 1 mg of isotopically clean 7 2 Ge, 7 4 G e or 7 6 Ge) is added to the solution. This allows the extraction efficiency to be monitored at the next extraction, at the end of the exposure, which typically lasts three weeks. After the extraction, the GeCl 4 swept out of the solution is reabsorbed in water and, after several stages of volume reduction, is converted to germane gas Ge H 4 . After purification the germane gas is introduced into a proportional counter of ~ 1 cm 3 volume (90% active). To optimize counter performance, old Xe is added, and the detector is operated at a slight overpressure. Very low background counters, made from very clean materials, have been developed for the experiment. They are operated in the G r a n Sasso lab itself, inside a P b - C u shielding. Some are surrounded by a N a l anti-counter. However, the N a l counters eliminate about as much background as they create, so that the global background level is essentially the same for the proportional counters with and without them. In these proportional detectors, one looks for the Auger electrons or X-rays emitted after electron capture in 7 1 Ge. Both Κ (10.37 keV) and L capture (1.17 keV) can be observed. The energy resolution is 26% F W H M f o r the Κ peak, and 4 3 % for the L peak. The combined acceptance for the Κ and L peaks in 2 F W H M windows is around 65.8%. Rise time cuts are also applied to further reduce the background. The corresponding acceptance is 97.7% for the L peak, and 95% for the Κ peak.

Neutrino Detectors

299

A mass spectroscopic analysis of the counting gas shows that typically 99% of the Ge is extracted. Around 90% ends up in the counter. In an additional test of the efficiency, 71 Ge was produced in the solution by electron capture in 7 1 As. It was shown that 99.8 + 3.7% of the atoms produced were extracted. As an ultimate test of the entire apparatus, a 62 PBq 5 1 Cr source emitting primarily 746 keV neutrinos was inserted into the tank. The ratio of measured to expected event rate was found to be 1.04 ±0.12. Counting with each sample of G e H 4 lasts for several half lives. Various time cuts are applied to reduce the background from radon gas inside and outside the detector. Thirty exposure of around 3 to 4 weeks were carried out in the period 1991-1993, corresponding to a total of 730 days. The Κ and L peaks are clearly visible. The count rate of the Κ and L candidate events in function of counting time shows a clear exponential spectrum, with the decay constant of 7 1 Ge, on top of a background, essentially flat, except for a small time dependent component due to 6 8 Ge (7j / 2 = 275 d). This 6 8 Ge was cosmogenically activated before the gallium was moved underground. The background rate per detector is typically 0.17 count per day, while the signal rate is around 0.33 per day, when counting starts. Taking into account the efficiencies, and after subtraction of 8 SNU to account for some known contributions not related to solar neutrinos, the main one being muon induced 7 1 Ge production, the solar neutrino associated 7 1 Ge production rate is 79 + 10 (stat.) + 6 (syst.) SNU (1σ). This correspond to a total of χ 136 observed solar neutrinos. This measured rate is less than that expected from calculations: (132 + 7) SNU [7], (123 + 7) [8]. Although smaller than that observed for higher energy neutrinos in the 37 C1 experiment, the discrepancy seems real, confirming the existence of a problem associated with solar neutrinos. Another radiochemical 7 1 G a experiment, SAGE [23], is in progress in the Baksan (Caucasus) underground lab, at a depth of 4700 mwe. The main difference is that the Ga is in metallic form, heated so as to be liquid (the melting point of Ga is 30°C). A specific chemical extraction procedure has been developed. A first series of 15 exposures of about 4 weeks each was performed, initially with 30 t of Ga, and now with 57 tons, contained in 8 tanks. Before each exposure, 700 μg of natural Ge is mixed to the Ga, in order to monitor the efficiency of the extraction procedure, typically 80%. As in GALLEX the extracted Ge is brought into the form of germane gas G e H 4 and introduced into a proportional counter. Some Xe is also added to improve the performance. For background reasons, only the Κ peak is presently used to search for 71 Ge. So far a total of %41 neutrinos has been observed above background, corresponding to 73 + 18/ — 16(stat.) + 5/ — 7(syst.) SNU. This confirms, with larger statistical errors, the GALLEX result. The experiment is continuing, and a calibration with a 5 1 Cr source has also been performed.

B. Direct measuring techniques in solar ν experiments 1. The Kamiokande-II

water Cerenkov

detector

Finally, solar neutrinos have been observed with the Kamiokande II water Cerenkov detector [19], designed originally to search for proton decay. The detection reaction

300

Jean-Luc Vuilleumier and Gabriele Zacek

is v e 4- e~ - > v e + e ~ . T h e d e t e c t o r is of cylinderical s h a p e , a n d c o n t a i n s 2 1 4 0 1 of w a t e r . T h i s v o l u m e is viewed b y a n a r r a y of 20 in. p h o t o - m u l t i p l i e r t u b e s , o n a 1 χ 1 m 2 lattice, o n t h e e n t i r e surface. T h e p h o t o - m u l t i p l i e r c o v e r a g e is 2 0 % of t h e t o t a l surface. T h a n k s t o a n efficient w a t e r p u r i f i c a t i o n system, a light t r a n s m i s s i o n l e n g t h of 48 m w a s a c h i e v e d . T h e i n n e r d e t e c t o r is c o m p l e t e l y s u r r o u n d e d b y a w a t e r C e r e n k o v d e t e c t o r w i t h a t h i c k n e s s of m o r e t h a n 1.4 m, t o identify c o n t a i n e d events, t o a b s o r b γ r a y s f r o m t h e surrounding rocks, a n d to monitor cosmic muons. Initially, in t h e s e a r c h f o r s o l a r n e u t r i n o s , t h e d e t e c t o r w a s t r i g g e r e d b y a t least 20 hit P M T ' s w i t h i n 100 ns. T h e trigger a c c e p t e d 7.6 M e V e l e c t r o n s w i t h 5 0 % efficiency, a n d 10 M e V e l e c t r o n s with 9 0 % efficiency. In t o t a l 4 5 0 d a y s of d a t a w e r e t a k e n w i t h these, o r slightly i m p r o v e d , settings, in 1 9 8 7 - 1 9 8 8 . I n a later stage, t h e g a i n of t h e p h o t o m u l t i p l i e r s w a s raised, w h i c h g a v e a b e t t e r sensitivity. T h e 5 0 % efficiency w a s a c h i e v e d a t 6.1 M e V , a n d t h e e l e c t r o n e n e r g y t h r e s h o l d in t h e a n a l y s i s w a s l o w e r e d t o 7.5 M e V . D a t a w e r e a c c u m u l a t e d for 590 d a y s in this m o d e in t h e p e r i o d 1 9 8 8 - 1 9 9 0 . T h e e n e r g y c a l i b r a t i o n is p e r f o r m e d w i t h y r a y s a r o u n d 8 M e V f r o m t h e r e a c t i o n N i (n, y)Ni ( a 2 5 2 C f n e u t r o n s o u r c e in a nickel c a n is i n s e r t e d at r e g u l a r i n t e r v a l s i n t o t h e w a t e r t a n k ) , with e l e c t r o n s f r o m m u o n d e c a y , a n d with t h e β d e c a y s f r o m s p a l l a t i o n p r o d u c t s of c o s m i c r a y m u o n i n t e r a c t i o n s . T h e e n e r g y (rms) in f u n c t i o n of t h e e l e c t r o n e n e r g y £ e is 2 2 % / [ £ e / 1 0 M e V ] 1 / 2 . T h e s p a t i a l i n f o r m a t i o n is r e c o n s t r u c t e d f r o m t h e t i m e of a r r i v a l of t h e light o n t h e p h o t o - m u l t i p l i e r s . T h e r m s a n g u l a r r e s o l u t i o n is 28° f o r a n e l e c t r o n of 10 M e V , t h e v e r t e x r e s o l u t i o n b e i n g 1.7 m . T h a n k s t o t h e high p u r i t y of t h e w a t e r , t h e b a c k g r o u n d f r o m active c o n t a m i n a n t s in it r e m a i n e d at a n a c c e p t a b l e level. T o r e d u c e t h e γ b a c k g r o u n d f r o m t h e r o c k s , o n l y e v e n t s f r o m a c e n t r a l fiducial v o l u m e of 6 4 0 1 a r e a c c e p t e d . P l o t t i n g t h e n u m b e r of e v e n t s in f u n c t i o n of t h e a n g l e t o t h e s u n 6>e, o n e sees a clear excess of e l e c t r o n s c o m i n g f r o m the s u n . T h e signal t o b a c k g r o u n d r a t i o is a r o u n d 0.5, a n d the t o t a l n u m b e r of e l e c t r o n s in t h e p e a k a t p + p + e

(£ l h = 1.44 MeV),

302

Jean-Luc Vuilleumier and Gabriele Zacek

in which case the electron can be detected, and its energy measured. This reaction has a 2 to 1 background to forward asymmetry with regards to the incident neutrino direction, which can thus be determined. Neutrinos of any flavor can interact by neutral-current interactions vx + d - > v x + p - l - n

(£ t h = 2.22 MeV).

To see these events, about 0.25% of NaCl will be dessolved in the inner tank. After thermalization, the neutrons will be captured on CI nuclei, releasing ys with 8.6 MeV of total energy. The C o m p t o n electrons of these ys will be detected with good efficiency. But of course, no information on the neutrino direction can be obtained in this way. In addition, neutrinos can scatter elastically off electrons: vx + e ~ - > v x + e - . Here the cross-section depends on the flavors of the neutrino. In this case again, the energy of the electron can be measured, as well as its direction, which is essentially that of the neutrino. Considering the energy thresholds, the experiment is essentially sensitive to 8 B neutrinos only. But the comparison of the event rates for the neutral and charged current reactions should provide a crucial, essentially model independent, test of neutrino flavor oscillations. It must be added that the experiment is very challenging. The total event rate is expected to be around 10 per day. To see so few events in such a large volume is of course not easy, and the background problems to be solved are numerous. All materials used must be radiochemically extremely clean. For instance, the heavy water must have a contamination in Th of less than 1 0 " 1 4 g/g! Construction is under way, and completion is expected in 1995.

2. BOREXINO,

a scintillation

detector

A scintillation detector, B O R E X I N O , is being developed and will be operated in the G r a n Sasso underground laboratory [82]. It consists of an inner spherical vessel of 4.2 m radius made from acrylic and filled with liquid scintillator. It is immersed in a spherical steel tank of 8 m radius filled with water. About 1600 photo-multipliers coupled to light guides are positioned in the water, at 2 m from the acrylic vessel. This allows a low threshold (250 keV), gives a good energy resolution (AE/E = 20%) and vertex resolution (13 cm at 0.5 MeV). The experiments will mostly see recoil electrons from v - e scattering, around 50 per day in a fiducial volume of 3 m radius (1001) assuming the full standard solar model flux, neutrinos from 7 Be contributing the most. Here also, the background problems from natural activities are enormous, and the radiopurities required are impressive. The U and Th concentration in the scintillator must be less than 10" 1 5 —10" 1 6 g/g, which apparently can be achieved.

Neutrino Detectors 3. New

303

techniques

F u r t h e r solar n e u t r i n o d a t a m a y c o m e f r o m I C A R U S , a 50001 liquid a r g o n i o n i z a t i o n d e t e c t o r with excellent spatial resolution, being considered for t h e G r a n Sasso u n d e r g r o u n d lab. N e u t r i n o s w o u l d be detected t h r o u g h elastic scattering off electrons, a n d charged c u r r e n t r e a c t i o n s o n a r g o n . O t h e r detection schemes, for instance using cryogenic detectors, are u n d e r s t u d y [ 6 ] . Particularly interesting a m o n g these is the p r o p o s a l [ 8 3 ] t o detect solar n e u t r i n o s using the reaction ve +

81

Br->e~ + 8 1 K r *

leading to the 1/2- excited state of 8 1 K r at 190 keV, with a lifetime of 13 seconds. T h e threshold is at 471 keV, a n d the experiment is sensitive t o t h e m o n o e n e r g e t i c 7 B e (higher line) a n d p e p neutrinos. T h e a b u n d a n c e of 8 1 B r is 4 9 . 3 % . T h e envisioned detector is a n a r r a y of 10 5 N a B r crystals, of 1 kilo each, o p e r a t e d in a b o l o m e t r i c m o d e at a t e m p e r a t u r e of 10 m K . T h e heat capacity is then so low t h a t t h e energy d e p o s i t e d in a crystal c a n be d e d u c e d f r o m the increase in t e m p e r a t u r e m e a s u r e d with a t h e r m i s t o r . G o o d c a n d i d a t e events a r e electron pulses, followed within o n e o r t w o lifetimes by t h e 190 keV deexcitation y ray pulse. T h e r e q u i r e m e n t of this coincidence will p r o v i d e a n efficient b a c k g r o u n d suppression, considering the excellent energy resolution achievable with b o l o m e t e r s . It will also m a k e it possible t o m e a s u r e with g o o d precision t h e energy of t h e electron, a n d thus t h a t of the i n c o m i n g n e u t r i n o . T h e p e a k s f r o m 7 Be a n d p e p n e u t r i n o s s h o u l d clearly stick out. Studies a r e u n d e r w a y t o establish t h e feasibility of this very delicate b u t powerful experiment.

B. Search for the magnetic moment of the ve at nuclear reactors A new d e t e c t o r for the s t u d y of v e e~ scattering is presently being developed [84], It consists of a 1 m 3 time p r o j e c t i o n c h a m b e r ( T P C ) filled with C F 4 at 5 bars. C F 4 is a low Ζ gas, with relatively low multiple scattering, a n d high density. T h e t o t a l n u m b e r of target electrons will be N e = 5.29· 10 2 7 . P u r i f i c a t i o n schemes exist which allow elect r o n s to be drifted over long distances, several meters. T h e cylindrical drift v o l u m e is delimited by an acrylic vessel, which s u p p o r t s t h e c a t h o d e on o n e end, a n d t h e grid, a n o d e a n d p i c k - u p p l a n e on the o t h e r one. T h e c a t h o d e a n d t h e grid define, with field shaping rings, t h e electric drift field. T h e signals a r e amplified a r o u n d t h e a n o d e wires. T h e p i c k - u p p l a n e b e h i n d t h e a n o d e measures t h e i n d u c e d signals, a n d p r o v i d e s t h e spatial i n f o r m a t i o n in t h e plane p e r p e n d i c u l a r t o t h e drift field. T h e spatial i n f o r m a t i o n in the direction parallel t o the drift field is d e t e r m i n e d f r o m t h e time e v o l u t i o n of t h e signals. T o reduce the b a c k g r o u n d , the T P C acrylic vessel is i m m e r s e d in a liquid scintillation detector, acting as shielding a n d a n t i - C o m p t o n c o u n t e r . T h e thickness of t h e scintillator is 50 cm. Also, materials with a low level of activity will be selected f o r t h e construction. T h e scintillator itself is s u r r o u n d e d by a P b layer t o p r o t e c t against activities f r o m t h e concrete. G o o d event c a n d i d a t e s will be single c o n t a i n e d electron

304

Jean-Luc Vuilleumier and Gabriele Zacek

tracks, coming roughly from the reactor core, with nothing in the anti-counter. Ends of tracks can be recognized from the increased ionization. It is planned to install this detector at the 2800 MWth Bugey reactor, at a distance of 18.6 m from the core. It is hoped that the background will be sufficiently low so that the energy threshold on the recoil electron can be lowered down to 500 keV. With that the event rate is expected to be around 10 per day, and the experiment will be sensitive to neutrino magnetic moments of order 2 - 3 · 10" 1 1 Bohr magnetons, about an order of magnitude better than in present experiments.

References [1] Schreckenbach, Κ., et al, Determination of the antineutrino spectrum from 2 3 5 U thermal neutron fission products up to 9.5 MeV, Phys. Lett. B, 160, 325, 1985. [2] Hahn, Α. Α., et al, Antineutrino spectra from 2 4 1 Pu and 2 3 9 Pu thermal neutron fission products, Phys. Lett. B, 218, 365, 1989. [3] Vogel P., Reactor antineutrino spectra and their application to antineutrino induced reactions, Phys. Rev. C, 24, 1543,1981. [4] Klapdor, Η. V., Antineutrino spectrum from the fission products of 2 3 9 Pu, Phys. Rev. Lett., 48, 127, 1982. [5] Zacek, G., et al, Neutrino-oscillation experiments at the Gösgen nuclear power reactor, Phys. Rev. D, 34, 261,1986. [6] Bahcall, J. N., Neutrino Astrophysics, Cambridge Univ. Press, Cambridge 1989. [7] Bahcall, J. N. and Ulrich, R., Solar model, neutrino experiments and helioseismology, Rev. Mod. Phys. 60(88)297; Bahcall, J. N. and Pinsonneault, Μ. H., Standard solar models, with and without helium diffusion, and the solar neutrino problem, Rev. Mod. Phys., 64, 885, 1992. [8] Turck-Chieze, S., et al, Revisiting the standard solar model, Astrophys. Jour., 335, 415, 1988; Turck-Chieze, S., and Lopes, I., Toward a unified classical model of the sun: on the sensitivity of neutrinos and helioseimology to the microscopic physics, Astrophys. Jour., 408, 347, 1993. [9] Turke-Chieze, S., et al, Comments on solar modeling, Nucl. Phys. Β (Proc. Suppl.), 28A, 116, 1992. [10] Hillebrandt, W., Neutrinos from collapsing stars and SN1987A, in Proc. Int. Workshop of Neutrino Physics, Klapdor, Η. V. and Pohv, B., Eds., Springer Verlag, Heidelberg 1987, 280. [11] Haxton, W. C., 3 7 Ar as a calibration source for solar neutrino detectors, Phys. Rev. C, 38, 2474, 1988. [12] Pontecorvo, B., Inverse beta processes and nonconservation of lepton charge, Sov. Phys. JEPT, 7, 172, 1958, [Zh. Eksp. Theo. Fiz., 34, 247, 1958]. [13] Bouchez, J., et al, Matter effects for solar neutrino oscillations, Z. Phys. C, 32,499, 1986. [14] Bilenky, S. M. and Petcov, S. T., Massive neutrinos and neutrino oscillations, Rev. Mod. Phys., 59, 671, 1987. [ 15] Mikeheyev, S. P., and Smirno v, A. Yu., Resonance enhancement of oscillations in matter and solar neutrino spectroscopy, Sov. J. Nucl. Phys., 42, 913, 1985, [Yad. Fiz, 42, 1442, 1985], [16] Boehm, F., and Vogel, P., Physics of Massive Neutrinos, 2nd edition, Cambridge Univ. Press, Cambridge 1992. [17] Boehm, F., et al, A large, low-energy neutrino detector for neutrino oscillations and Supernovae watch, Nucl. Instr. Meth. A, 300, 395,1991. [18] Rowley, J. K., Cleveland, Β. T. and Davis Jr., R., The chlorine solar neutrino experiment, in Solar neutrinos and neutrino astronomy, M. L. Cherry, W. A. Flower and K. Lande editors, (New York: American Institute to Physics) Conf. Proc. 126, p.l, 1985.

N e u t r i n o Detectors

305

[19] Hirata, Κ. S., et al., Observation of 8 B solar neutrinos in the Kamiokande-II detector, Phys. Rev. Lett. 63, 16, 1989; Hirata, Κ. S., et al., Results from one thousand days of real time, directional solar neutrino data, Phys. Rev. Lett. 65, 1297, 1990; Hirata, K. S., et al., Constraints on neutrino oscillations parameters from the Kamiokande-II solar neutrino data, Phys. Rev. Lett. 65,1301,1990; Hirata, K. S., et al., Real-time, directional measurement of 8 B solar neutrinos in the Kamiokande II detector, Phys. Rev. D, 44, 2241, 1991. [20] N a k a m u r a , K., Recent results from K a m i o k a n d e solar neutrino observation, in Proc. N E U T R I N O 92, Nucl. Phys. B(Proc. Suppl.) 31, 105, 1993. [21] Totsuka, Y., Solar neutrinos: real-time experiments, in Relativistic Astrophysics and Particle Cosmology, Proc. T e x a s / P A S C O S 92, Annals of the N e w York Academy of Sciences, Vol. 688, 344, 1993. [22] Anselmann, P., et al., Solar neutrinos observed by G A L L E X at G r a n Sasso, Phys. Lett. B, 285, 376, 1992; Anselmann, P., et al., Implications of the G A L L E X determination of the solar neutrino flux, Phys. Lett. B, 285, 390, 1992; Anselmann, P., et al., G A L L E X results from the first 30 solar neutrino runs, Phys. lett. B, to appear in 1994. [23] Abazov, Α. I., et al., Search for neutrinos from the sun using the reaction 7 1 Ga(v e ,e) 7 1 Ge, Phys. Rev. Lett., 67, 3332, 1991; Abdurashitov, J. N., et al., Results f r o m SAGE, (the Russian-American gallium solar neutrino experiment), Phys. Lett. B, to appear in 1994. [24] Cribier, M., et al., Study of a high intensity 746 keV neutrino source for the calibration of solar neutrino detectors, Nucl. Instr. Meth. A, 265, 574, 1988. [25] Lee, B. W., and Shrock, R. E., N a t u r a l suppression of symmetry violation in gauge theories: M u o n and electron number nonconservation, Phys. Rev. D16, 1444, 1977. [26] Marciano, W., and Sanda, A. I., Exotic decays of the M u o n and heavy leptons in gauge theories, Phys. Lett. B, 67, 303, 1977. [27] Voloshin, Μ. B., O n the compatibility of a small mass and large magnetic m o m e n t of the neutrino, Sov. J. Nucl. Phys. 48, 512, 1988 [Yad. F i z , 48, 804, 1988. [28] Babu, K. S. and M o h a p a t r a , R. N., Models for large transition magnetic m o m e n t of the electron neutrino, Phys. Rev. Lett. 63, 228, 1989. [29] Leurer, M. and Marcus, N., A model for large neutrino magnetic transition m o m e n t and naturally small mass, Phys. Lett. B, 237, 81, 1990. [30] Barbieri, R., et al., Supersymmetry, R-parity breaking and the neutrino magnetic moment, Phys. Lett. B, 252, 252, 1990. [31] Barr, S. M., Friere, Ε. Μ., and Zee, Α., Mechanism for large neutrino magnetic m o m e n t , Phys. Rev. Lett., 65, 2626, 1990. [32] Pal, Β. K., Zee model on M a j o r a n a neutrino mass and magnetic moment, Phys. Rev. D 44, 2261,91. [33] Vogel, P. and Engel, J., Neutrino electromagnetic form factors, Phys. Rev. D, 39, 3378, 1989. [34] Kayser, B., et al., Charged and neutral-current interference in v e - e scattering, Phys. Rev. D, 20, 87, 1979. [35] Lande, Κ., talk given at Neutrino 92, the 15th Int. Conf. on N e u t r i n o Physics and Astrophysics, G r a n a d a , Spain, 7 - 1 2 June 1992. [36] Voloshin, Μ. B., Vysotskii, Μ. I., and O k u n , L. B., Neutrino electrodynamics and possible consequences for solar neutrinos, Sov. Phys. J E T P 64, 446, 1986 [Zh. Eksp. Teor. Fiz. 91, 754, 1986], [37] Lim, C. S. and Marciano, W. J., Resonant spin-flavor precession of solar and supernova neutrinos, Phys. Rev. D 37, 1368, 1988. [38] Akhmedov, Ε. K., Resonant amplification of neutrino spin rotation in matter and the solar-neutrino problem, Phys. Lett. B, 213, 64, 1988. [39] Babu, K. S., M o h a p a t r a , R. N. and Rothstein, I. Z., Reconciling the time dependence of the Chlorine and K a m i o k a n d e solar-neutrino data, Phys. Rev. D 44, 2265, 1991. [40] Ono, Y., and Suematsu, D., The role of low energy neutrinos in the solar neutrino problem, Phys. Lett. B, 271, 165, 1991. [41] Akhmedov, E. Kh., Lanza, A. and Pectov, S. P., Implications of gallium solar neutrino d a t a for the resonant spin-flavor precession scenario, Phys. Lett. B, 303, 85, 1993.

306

Jean-Luc Vuilleumier and Gabriele Zacek

[42] Berezhiani, Z. G., et al., Neutrino oscillations and magnetic moment transitions in a model with a conserved lepton number, Phys. Lett. B, 264, 381, 1991. [43] Lattimer, J. M. and Cooperstein, J., Limits on the neutrino magnetic moment from SN 1987A, Phys. Rev. Lett. 61, 23, 1988. [44] Barbieri, R. and M o h a p a t r a , R. N., Limit on the magnetic m o m e n t of the neutrino from Supernova 1987A observations, Phys. Rev. Lett., 61, 27, 1988. [45] Notzold, D., New bounds on the neutrino magnetic moments from stellar collapse, Phys. Rev. D, 38,1658, 1988. [46] Bernstein, J., et al., Electromagnetic properties of the neutrino, Phys. Rev., 132, 1227,1963. [47] Sutherland, P., et. al., Astrophysical limitations on possible tensor contributions to weak neutral interactions, Phys. Rev. D, 13, 2700, 1976. [48] Raffelt, G., New b o u n d on neutrino dipole m o m e n t s from globular clusters stars, Phys. Rev. Lett., 64, 2856, 1990. [49] Castellani, V. and Degl'Innocenti, S., Stellar evolution as a probe of neutrino properties, Astrophys. Jour., 402, 574, 1993. [50] Yamaguchi, H., Interactions induced by high energy neutrinos, Prog. Theo. Phys. A, 377, 474,1982. [51] Particle d a t a group, Review of Particle Properties, Phys. Rev. D, 45, VII. 8, 1992. [52] Wilkinson, D. H., Analysis of neutron ß-decay, Nucl. Phys. A, 377, 474, 1982. [53] Vogel, P., Analysis of the antineutrino capture on protons, Phys. Rev. D, 29, 1918, 1984. [54] Fayans, S.A., Radiative corrections and recoil effects in the reaction v e + ρ -> η + e + at lo w energies, Sov. J. Nucl. Phys., 42(4), 590, 1985, [Yad. Fiz., 42, 929, 1985], [55] Ali, A. and Dominguez, C. Α., Neutrino (antineutrino) disintegration of the deutron and the structure of the weak neutral current, Phys. Rev. D, 12, 3673,1975. [56] Reines, F., Sobel, Η. W. and Pasierb, E., Evidence for neutrino instability, Phys. Rev. Lett., 45, 1307, 1980; Pasierb, E. et al, Detection of weak neutral current using fission v e on deuterons, Phys. Rev. Lett., 43, 96, 1979. [57] Ewan, G. T., et al., Sudbury neutrino observatory proposal, SNO-87-12, 1987. [58] Perkins, D. H., Introduction to High Energy Physics, 3rd edition, Addison-Wesley Publ. Comp. Inc., 1987. [59] Reines, F. and Cowan, C. L., Detection of free neutrinos, Phys. Rev., 92, 830, 1953. [60] Cowan, C. L. et al., Detection of the free neutrino: a confirmation, Science, 124,103,1956. [61] Reines, F. and Cowan, C. L., Free antineutrino absorption cross section. I. Measurement of the free antineutrino absorption cross section by protons, Phys. Rev., 113, 273, 1959. [62] Nezrick, F. A. and Reines, F., Fission-antineutrino interaction with protons, Phys. Rev., 142, 852, 1966. [63] Kwon, H., et al., Search for neutrino oscillations at a fission reactor, Phys. Rev. D, 24,1097,1981. [64] Cavaignac, J. F., et al. Indication for neutrino oscillations from a high statistics experiment at the Bugey reactor, Phys. Lett. B, 148, 387, 1984. [65] de Kerret, H., New results of neutrino oscillation search at the Bugey reactor, in Proc. 23rd. M o r i o n d W o r k s h o p , Fackler.O. and T h a n h Van, J. T., Eds., editions Frontieres, 125,1988. [66] Zacek, G., Search for neutrino oscillations at a fission reactor, thesis Technische Universität München, München 1984. [67] Afonin, Α. Ι. et. al., Cross section for the inverse beta decay of the p r o t o n from nuclearreactor measurements, J E T P Lett., 38(7), 436, 1983, [Pis'ma Zh. Eksp. Theo. Fiz., 38(7), 361,1983], [68] Afonin, Α. I. et al., Search for neutrino oscillations in an experiment in the reactor of the Rovno nuclear power plant, J E T P Lett., 42(5), 285, 1985, [Pis'ma Zh. Eksp. Theo. Fiz., 42(5), 230, 1985], [69] Belenki, S. N. et al., New experiment on the inverse β decay of the proton in a nuclear reactor, J E T P Lett., 38(8), 493,1984, [Pis'ma Zh. Eksp. Theo. Fiz., 38(8), 406, 1983], [70] Belenki, S. N. et al., Measurement of the cross section for inverse β decay by a new method in a reactor experiment at the Rovno nuclear power station, Sov. J. Nucl. Phys. 40(1), 93, 1984, [ Y a d . Fiz., 40, 148, 1984],

Neutrino Detectors

307

[71] Vidyakin, G. S. et al., Detection of antineutrinos in the flux from two reactors, Sov. J. Nucl. Phys., 66(2), 243, 1987, [Zh. Eksp. Theo. Fiz., 93,424, 1987], [72] Vidyakin, G. S., et al, Bounds on the neutrino oscillation parameters for reactor neutrinos, Sov. Phys. JEPT, 71(3), 424, 1990, [Zh. Eksp. Theo. Fiz., 98, 764,1990], [73] Collot, J., A new generation detector of low-energy anti-neutrinos, in Proc. 23rd. Moriond Workshop, Fackler, O., and Thanh Van, J., T., Eds., Edition Frontieres, 153, 1988. [74] Avenier, M., Neutrino oscillation search at the Bugey reactors, in Proc. of the Moriond workshop, Fackler, O., Fontaine, G., Thanh Van, J. T., Eds., editions Frontieres, 111,1991. [75] Reines, F., Do Neutrinos oscillate? Nucl. Phys. A, 306, 469, 1983. [76] Davis, B. R., et al, Reactor antineutrino spectra and their application to neutrino induced reactions, Phys. Rev. C, 19, 2259, 1979. [77] Vidyakin, G. S., et al., Observation of weak neutral current in interactions of fission antineutrinos with deuterons, JETP Lett., 51, 279, 1990. [78] Reines, F., Gurr, Η. S., and Sobel, Η. W., Detection of ve — e scattering, Phys. Rev. Lett., 37, 315, 1976. [79] Avignone, F. T., V - A elastic scattering of electrons by fission antineutrinos, Phys. Rev. D, 2, 2609, 1970. [80] Geiregat, D., et al, An improved determination of the electroweak mixing angle from muon-neutrino electron scattering, Phys. Lett. B, 259, 499, 1991. [81] Gurevitch, 1.1., et al, Study of reactor antineutrino scattering by electron, using organofluoric based detector, preprint, I.V. Kurtchatov Institute of Atomic Energy, 123182 Moscow, Russia, 1991; Vidyakin G. S., et al, Study of the scattering of fission antineutrinos by electrons with an organofluoric scintillator detector, J E T P Lett., 49, 740, 1989. [82] Borexino at Gran Sasso: Proposal for a real time detector for low energy solar neutrinos. [83] Alessandrello, Α., et al, A cryogenic experiment for solar neutrino spectroscopy and search for dark matter, preprint INFN/AE-92/28, University of Milan, 1992. [84] M U N U collaboration, proposal; Broggini, C. et al, A gas detector to measure the ve magnetic moment at a nuclear reactor, Nucl. Inst, and Meth. A 311, 319, 1992.

9

J. P.

Fragment Multi-Detector for the Study of Hot and Dense Nuclear Matter Produced in Relativistic Heavy-Ion Reactions Coffin

Table of Contents I. Introduction 310 II. Basis of the detector design 312 A. Physics consideration 312 B. Basic concepts for the design of the detector 313 C. Basic solutions 314 III. Detector 315 A. General description 315 B. T h e target area setup 316 1. T h e start/anti halo/reaction counter system 316 2. Beams and targets 318 C. T h e plastic scintillator wall 318 1. T h e inner wall 318 2. The outer wall 320 D. T h e cluster detector ensemble 323 1. The rosace 324 2. The parabola 326 E. The cluster detector-scintillator wall association 328 1. Multi-hit treatment and tracking 328 2. Performances 330 IV. Laser system and calibration procedures 332 A. Laser and light distribution system 332 B. Laser operation and surveillance 333 C. Calibration procedures 334 V. Electronics and d a t a acquisition 335 A. General scheme 335 B. Trigger system 336 C. D a t a acquisition 337 VI. Acceptance and thresholds 337 VII. Background 339 VIII. The Phase II of the detector 340 IX. Summary and outlook 341 References 341

310

Jean-Pierre Coffin

I. Introduction The production of hot and dense nuclear matter in collisions of relativistic heavy-ions was first proposed by Sheid, Müller and Greiner [1] in 1974. This strongly compressed state of the nuclear medium was supposed to be created by propagation of a shock wave from the impact zone between the colliding nuclei. This was not the first suggestion that shock waves exist in nuclear reactions since they had already been considered by other authors [2-4]. However, it was the first time that such a mechanism was viewed as being capable of producing hadronic matter at a density several times higher than normal (0.16 nucleon/fm 3 ) and at a temperature of the order of a hundred of MeV. Thus a new field of research was proposed allowing to some extent the production and the study of nuclear matter as it exists in black holes, neutron stars, exploding super-novae..., hence opening up a window to outer space and on matters relating ultimately to the origin of the universe. The physics of relativistic heavy-ions was born. Schematically these reactions can be described as follows: In the first stage the excited zone formed in the collision grows and propagates from the point of contact between projectile and target as a consequence of a shock wave front. In such a process, compression energy and thermal energy are produced, the latter resulting from nucleon-nucleon collisions whose frequency rises as the mean-free path is reduced; in the meantime hadronic resonances (Δ, N*) are formed. This phenomenon, which brings the system to a stage of high temperature and density and large internal energy, lasts only a very short time (t ~ 20fm/c). In the second stage hard photons, electron pairs, and mesons created in the free space and from resonance decay are promptly emitted while the nucleonic system expands and disassembles to generate light particles (n, p, d, t, 3 , 4 He) and fragments whose mass numbers are in the A = 4-30 range. During this phase the potential compression energy is mostly converted into kinetic energy and communicates a collective motion to the bulk of particles and fragments. Due essentially to weak viscosity effects, only a little compression energy is transformed into thermal energy. Also, during this phase (ί ~ 40 fm/c) the temperature drops considerably. At last, in the final stage, the system expands some more and reaches a dilute density state, compared to the ground-state density, where the fragments cease to interact with each other. From this point on, only the unstable fragments may still particle decay. This whole particle and fragment emission is often placed under the generic name of multifragmentation. Its nature and yield depend much on the incident energy. As the latter increases, the system tends to disassemble into a growing number of fragments of diminishing mass. As long as the projectile energy is lower than ~ 700 A MeV, multifragmentation remains essentially a one decay process because the mesonic component formed in the early stage of the reaction is then practically nonexistent. The above description of relativistic heavy-ion reactions applies, of course, to the "participant" zone of the nuclear system corresponding geometrically to the region where the projectile and target are in contact. There are consequently also "spectator" parts corresponding to the non-directly interacting fractions of the projectile and target. The lower the reaction's impact parameter, the larger the "participant" zone. Thus, a relevant and comprehensive study of relativistic heavy-ion reaction mechanisms and of the properties of hot and dense nuclear matter, in the incident energy

Fragment Multi-Detectors

311

domain spanning 100 to 600,4MeV, implies the measurement of neutrons, light charged particles and fragments emitted from highly central collisions. The ambition of any modern and effective experiment should consist of measuring all these products simultaneously. However, it immediately becomes obvious that by their very nature neutrons have to be studied in an instrumental context, different from that of charged particles and fragments. Let us look for an instant at the early experimental attempts in this field. In spite of theoretical guidance, the experimentalists had to face a great deal of difficulty because they had practically no basis for comparison or extrapolation. A first generation of studies relying upon emulsions [5], streamer chambers [6], forward angle spectrometers [7, 8], and solid state detectors [ 9 - 1 2 ] pioneered the field from the late seventies to the early eighties. It quickly became obvious that sophisticated a p p a r a t u s was imperative for investigating this complicated physics. In today's standard experiment, it is necessary to approach conditions as closely as possible where measurements are performed event-by-event over the full solid angle. All charged particles and fragments from each single event have to be A (or atomic number Z) identified as well as having their velocity or linear m o m e n t u m determined over a wide dynamic range. Such an ambition led to the generation of the so-called 4n-detector systems whose Plastic Ball [ 1 3 - 1 5 ] and streamer chamber [ 1 4 - 1 6 ] at the Bevalac at Berkeley (USA) and Diogene [17] at Saturne in Saclay (France) were leading examples. They illustrate the various technical solutions which have been formulated so far. The Plastic Ball combines AE-E and ΔΕ-time-of-flight measurements to yield A and Ζ determination, respectively. It covers 9 7 % of 4π and allows very good particle identification for hydrogen and helium isotopes and also π + detection with high efficiency. It has been responsible for the vast and extremely rich harvest of experimental results from relativistic heavy-ion reactions collected during the eighties. It was more recently moved to C E R N (Geneva) with the aim of participating in experiments, using ultra-relativistic 1 6 0 and 3 2 S heavy-ion beams, directed to the search for signals of quarkgluon plasma formation. The main limitations of this apparatus are its difficulty in stopping the emitted particles in the detector medium at incident energies exceeding 300 A MeV or so, an reliable fragment identification restricted to Ζ < 6, and a somewhat insufficient detector segmentation responsible at forward angles for non-negligible multiple hits in high emission rate reactions. The detector Diogene associates a forward part (Θ < 37°) very similar to that of the Plastic Ball, with an ensemble of drift chambers operating in the longitudinal magnetic field of a solenoid. This second part, covering the 37° < θ < 119° polar angle domain, allows particle mass identification together with their spatial distribution by means of magnetic analysis. The main advantage of this detector is that these tracking chambers can be triggered and gated. Thus high counting rates together with high beam intensities can be used to study rare phenomena. This apparatus was designed essentially for studying pions and the Ζ = 1 and Ζ = 2 isotopes with multiplicity less than ~ 35. Presently, besides these classes of detectors, forward magnetic spectrometers also exist, generally associated with electronic counters in their focal plane. The Aladin spectrometer [18] using a multiwire gas chamber [18,19] on its detection surface, which came into operation very recently at G.S.I., D a r m s t a d t (Germany), is a good

312

Jean-Pierre Coffin

example of such devices. These spectrometers are useful to study very forward oriented emissions, meaning those from asymmetric nuclear systems produced in so-called inverse kinematics reactions or those induced from peripheral collisions. Another type of detector is the multiwire gas chamber. Its main advantage resides in the fact that a large portion of the solid angle is covered with a single homogeneous detector which can be complemented by other devices. Such a design is illustrated by the Time-Projection-Chamber [20] placed in the HISS magnet at the Bevalac. The replacement of wires by pads and the enormous progress made in fast electronics and in their miniaturization have brought the performance of these detectors up to a high standard. The aim of this chapter is to furnish the reader with the state of the art report on charged particle and fragment detection in the energy range of one to several hundred MeV/nucleon and within a large solid angle. O n e possibility is to review all existing methods potentially capable of coping with this task. It would be a bit difficult to weight the merits and demerits of each method correctly if discussed in any actual and precise experimental context. Indeed, the universal counter does not exist and the best choice is strongly dependent on the specific detection one wants to emphasize. Another possibility is to examine in detail all classes of detectors dedicated to the detection of multifragmentation products. This would certainly require several hundred pages which is far beyond the scope of the present effort. In several cases, it would be of historical interest only. It is the conviction of the author that one effective approach is to present a recently constructed and satisfactorily performing apparatus which overall supersedes all previous realisations. We have chosen to present the 4π Facility FOPI [21] installed at G.S.I, as a comprehensive and illustrative example. Its description will consist first of explaining in some detail the m a j o r requirements that have to be fulfilled to give a high level apparatus. Secondly, the leading ideas underlying the technical choices made will be presented. Thirdly, the various parts of the detector will be described, and examples of the performance achieved will be shown.

II. Basis of the dectector design A. Physics considerations We have seen that relativistic reactions create hot and dense nuclear matter whose properties are governed partly by the amount of energy stored into the system. The energy is divided into approximately two equal fractions: a potential compression energy and a thermal energy. The former produces a collective motion of nucleons and clusters, the so-called collective flow which contains implicitly a great deal of information about the compressibility and the viscosity of nuclear matter. The latter generates the statistical decay of the system. This model is certainly oversimplified and a complete description would require a microscopic approach in which the dynamical evolution and the deexcitation of the system are intimately correlated. At present, three classes of theoretical models have been developed: microscopic, hydrodynamic and statistical. O n the whole, they all try to relate the microscopic one-body and two-body

Fragment Multi-Detectors

313

interactions to the macroscopic properties of nuclear matter, like the compressibility and viscosity that enter into the hydrodynamic description, as well as to the thermostatic characteristics like temperature and entropy. Such an enterprise aims ultimately to establish the equation of state of nuclear matter. In order for the experimentalist to proceed with stringent tests of the theoretical models, he has to characterize each event completely and independently by measuring the following main quantities for all its components: the atomic mass A or the atomic number Ζ of the particle, the velocity ν and the initial direction of emission (for instance in terms of polar 0 and azimuthal φ angles). From these primary observables associated variables can be deduced. These are the multiplicity Μ of each particle species, the rapidity y, the linear momentum ρ and the kinetic energy Ε of every particle. This set of variables defines consequently an ensemble of basic requirements which has to be incorporated in the design of the detector.

B. Basic concepts for the design of the detector In the light of the requirements set out in the previous paragraph it is necessary to design a detector featuring the following capabilities: 1) Simultaneous detection of all emitted products in a single event in a geometry as close as possible to 4π. This not only allows a complete characterisation of the event but also permits the study of correlations between components within an event. The decay products are photons, leptons, mesons, baryons and fragments. However, it is practically impossible to design a detector offering a dynamic range so that one can measure all of them at the same time. Restriction to a limited domain of decay products, for instance π 1 , charged particles and fragments, is inevitable. Furthermore, at projectile energies where fragment emission is substantial, the meson cross-section is very low and vice versa. Thus a detector dedicated to charged particles and fragments is only suitable for studying multifragmentation. 2) Azimuthal symmetry of the detector in order to determine the emission configuration of all components of a given event. This is very useful for analysing phenomena relating to the collective flow, for instance. 3) High granularity of the detector to deal with large multiplicities as encountered in head-on collisions of very heavy systems, like Au + Au, and to reduce the multiplehit confusion. 4) Good segmentation to allow precise determination of the spatial orientation (θ, φ) of the detected products and also to measure accurate correlations between particles within an event. 5) Accurate measurements of energy loss AE in a detector medium and of time-of-flight (ToF) τ, thus allowing good Ζ characterisation of the detected particles and fragments. Determination of A is, of course, preferable to a determination of ρ and E. When only Ζ is measured, one applies the A = 2Z recipe which is somewhat a limitation, at least for the Ζ = 1 and 2 isotopes. 6) Low experimental thresholds for the detection of the slowest particles.

314

Jean-Pierre Coffin

7) Versatility and high selectivity of triggers allowing a good sampling of the events, for instance to select collisions according to their centrality and also permitting the search of rare processes. 8) Advanced modularity for easier mounting, adjustment and repair of all parts. 9) Least possible shadowing by the various parts of the apparatus to minimise scattering and loss of charged particles and to offer high transparency to other reaction products, like neutrons, in provision of joint measurements. 10) Possible association of the detector with other devices devoted to measurements inaccessible to a charged particle detector: photons, π°, neutrons.

C. Basic solutions To cope with the requirements listed above one is led to the typical solutions which follow: 1) In the present state of the art symmetrical reaction systems are preferentially studied because of the convenience that they carry by nature. This is true also for theoretical calculations where A + A systems or systems including Ν = Z nuclei are often considered. Asymmetric systems generally require one step higher in complexity or specificity. Thus, at projectile energies exceeding 100 A MeV or so, a 0° < θ < 30° coverage in the laboratory system allows the quasi-complete forward-hemisphere detection in the centre-of-mass frame. As a first approach, detectors limited to this polar angle range with full azimuthal coverage are suitable. Of course, true 4π hermiticity in the lab frame is better and detectors offering this possibility should eventually be constructed. 2) At the bombarding energies considered in these studies, the particle multiplicities are very high at low angles (Θ < 1°) and the range of products to detect is wide (from protons to heavy projectile-like fragments). Furthermore, these products have to be separated from the primary ion-beam which must be damped somewhere. For these reasons, the detectors do not generally cover angles less than 1°. 3) In view of the large kinetic energies imputed to some of the emitted products, thick detectors are necessary for a Ε — τ measurements to allow A determination. However, large size counters are often more costly, more difficult to construct and to install than detectors devoted to ΔΕ measurements. They are also inappropriate for excitation functions due to the lack of versatility that these measurements demand in terms of dynamics. Thus ΑΕ — τ measurements are generally preferred. 4) ToF measurements are needed to determine Z. At the projectile energies considered, the particles and fragments to be detected have velocities typically in the 3-15 cm/ns range. Consequently, good ΔΖ/Ζ resolution imposes that τ be measured over a flight-path of about 3 m minimum. 5) A compromise between detector size and sufficient granularity for an affordable cost leads to a number of detection elements around 1000. In the present state of the art, this figure constrains the choice of these elementary devices to scintillator-phototube ensembles or ionisation chambers. It excludes expensive or so far low dimension detectors like solid-state Si counters.

Fragment Multi-Detectors

315

III. Detector A. General description The 4π detector FOPI built at G.S.I, attempts to feature all the capabilities described above that a modern apparatus should incorporate and to use the methods and solutions according to the rational developed in the previous section. It has been constructed within a large collaboration involving physicists primarily from France and Germany but also from several countries of eastern Europe. T w o steps can be distinguished in the design and construction of this detector. The first (Phase I) has been operational since mid-1990. The second has been in actual operation since the end of 1993. Phase I is devoted exclusively to multifragmentation studies and will be described here in some detail. Phase II extends largely the solid angle coverage and allows Ζ = 1 and 2 particle measurements, but the main emphasis is put on π f and Κ meson detection. For completeness, it will be presented briefly also. Phase I corresponds to an ensemble of detectors, covering the 1 0 < θ < 30° lab polar angle domain, for which the overall sizes are about 4 m in diameter and 3 m in depth. It is devoted to the detection of charged particles and fragments (1 < Ζ < 18), it comprises as illustrated in Fig. 1: - A Plastic Scintillator Wall subdivided into - an Inner Wall (IW) with 1 0 < θ < 7° - an Outer Wall (O W) with 7° < θ < 30° For mechanical reasons which will become clear later, the various elements of the wall

τ

Fig. 1. Schematic view of the Phase 1 of the 47i-Detector FOPI at G.S.I., D a r m s t a d t . T h e symbols indicate the following components: Τ = Target, H B = Helium Bag, Ρ = Parabola, R = Rosace, IW = Inner Wall, O W = O u t e r Wall. All these elements are described in the text.

316

Jean-Pierre Coffin

are not in a single plane but consist of a succession of shifted planes. T h u s the wall terminology is incorrect since the ensemble is cone shaped. It measures the energy loss AE a n d the time-of-flight τ, hence the velocity υ of the detected elements, to allow their Ζ determination. It is a d e q u a t e for particles a n d fragments 1 < Ζ < 9. - A cluster detector system, placed before the Plastic Wall. It is also subdivided into two components. - the Rosace (R), covering 1° < Θ < 6°, m a d e of thin plastic scintillators. - t h e P a r a b o l a (P), covering 6° < θ < 30°, constituted of 16 gas filled ionization chambers. It also yields AE information, which combined with τ given by the Plastic Wall, extends the identification of fragments towards higher Ζ (1 < Ζ < 18) a n d lower rapidity t h a n possible with the Plastic Wall.

B. The target area setup As in most of the experiments of this kind performed to date the target and the detection ensemble are placed in air. In view of the size a n d the complexity of the detector a n d its associated electronics, it is not conceivable to place the whole system under vacuum. An alternative solution would consist of keeping only the target environment under vacuum. T h e present solution might be viewed as a weakness at first glance, since particles a n d even m o r e critical fragments to be detected lose energy over the a b o u t 4 m of air separating the target from the detector. However, as it is now, only the b e a m ions but not the detected p r o d u c t s have to traverse a v a c u u m window which would generate a m u c h m o r e i m p o r t a n t energy loss for the fragments t h a n in the air. T h e d r a w b a c k is a certain rate of b a c k g r o u n d reactions on the b e a m exit window and in the along the b e a m path. W h e t h e r this is tolerable or not depends on the experimental conditions: target thickness, b e a m energy, range of expected f r a g m e n t Z, The present solution turns out to be the best c o m p r o m i s e in studies of fairly massive systems, like Au + Au, at relatively low b o m b a r d i n g energies, such as 100-400 A MeV, where f r a g m e n t a t i o n is d o m i n a n t . Let us now examine the different c o m p o n e n t s of the detection ensemble, beginning with the setup of the target area.

1. The start/and

halo/reaction

counter

system

As shown in Fig. 2, the b e a m leaves the under-vacuum section a b o u t 50 cm upstream of the target by passing t h r o u g h a 125 μηι K a p t o n window. Before hitting the target it goes first t h r o u g h a slit system consisting of 4 (up-down-left-right) scintillator paddles of 5 m m thickness, each coupled to a photomultiplier, used as a b e a m halo suppressor ( H J or a b e a m position control. Both horizontal and vertical openings can be changed by remote control; 9 x 9 m m 2 is a typical opening which vetoes a b o u t 1% of the incident b e a m and any change in its structure is readily seen. T h e n the b e a m goes t h r o u g h a device delivering an initial timing signal (t start ). It consists of a 50 μπι thick scintillator foil, tilted 45° to the b e a m axis, the emitted light

Fragment Multi-Detectors

317

-185 -280 •

-170-90 • 40-»U-S0-J D

-195-

- 1 1 0 -

PM ι

1 Beom

L l Target

v . c ΰ

tAO,

Π

5U

U

r D

D

Hi

Stort-box

H2

i*-60—ι Reoction Counter (Rc)

Fig. 2. Schematic view of the Start/Anti-halo/Reaction-counter system. The various components and their use are described in the text. Unless otherwise specified all distances are expressed in millimeters.

b e i n g p a r t l y collected w i t h t w o p h o t o m u l t i p l i e r s p l a c e d s y m m e t r i c a l l y w i t h r e s p e c t t o t h e b e a m axis. T h e sensitive p a r t s (scintillators a n d p h o t o c a t h o d e s ) a r e h o u s e d in a s o f t - i r o n b o x w h o s e e n t r a n c e a n d exit b e a m o p e n i n g s a r e closed w i t h t h i n p l a s t i c foils in o r d e r t o a v o i d p a r a s i t i c light interference. F r o m t h e t i m e signals ( t j a n d t 2 ) d e l i v e r e d b y these t w o t u b e s o n e g e n e r a t e s a start signal f o r t h e τ m e a s u r e m e n t s . T h i s l a t t e r is f r o m a c o i n c i d e n c e of t h e tl a n d t2 signals w i t h a v e t o f r o m T h e time difference — r 2 ) is a l s o r e c o r d e d s e p a r a t e l y . It c a n b e s u b t r a c t e d f r o m (or a d d e d t o ) t1 u s e d as start (or stop) signal. T h i s yields a n overall start signal (f s t a r t ) w h i c h is a f a c t o r 1 / y j 2 b e t t e r in r e s o l u t i o n t h a n if d e t e r m i n e d a g a i n s t t h e m e a n t i m e of t1 (this p o i n t is discussed f u r t h e r in Sect. III. C . 2). A so-called R e a c t i o n c o u n t e r (Rc) is installed n e a r a n d u p s t r e a m of t h e t a r g e t . Its f u n c t i o n is d e s i g n e d t o d e t e c t b a c k w a r d e m i t t e d p r o d u c t s f r o m t h e t a r g e t . It is s u p p o s e d t o m e a s u r e a t least o n e f r a g m e n t of a c e n t r a l i m p a c t collision in t h e t a r g e t , w h e r e a s t h e p r o b a b i l i t y of seeing r e a c t i o n s o n o t h e r m a t e r i a l s is m u c h lower. T h i s d e t e c t o r c o n s i s t s of a p s e u d o - r i n g scintillator of 30 m m i n n e r d i a m e t e r w h e r e t h e b e a m p a s s e s t h r o u g h its c e n t r a l hole. It is p r o t e c t e d f r o m particles p r o d u c e d in u p s t r e a m r e a c t i o n s b y a n o t h e r h a l o s u p p r e s s o r system (H 2 ), identical t o Η ( , p l a c e d in f r o n t of it. It is g e n e r a l l y o p e n e d t o 15 χ 15 m m 2 . T h e f r o n t e n d of H 2 f a c i n g t h e t a r g e t is p r o t e c t e d f r o m b a c k s c a t t e r i n g by a 1 m m l e a d foil. W i t h this s e t u p it is p o s s i b l e t o g e n e r a t e t h e f o l l o w i n g h a r d w a r e t r i g g e r signals: G o o d b e a m (GB) o r g o o d s t a r t J G S ) = ( t j - t ^ - H j B e t t e r b e a m (BB) = ( i 1 t 2 ) - H 1 H 2 M i n i m u m reaction (MR) = (tj •t2)-H1 - H 2 - R c

318

Jean-Pierre Coffin

2. Beams and

targets

The beams are delivered by the rapid cycling synchrotron SIS using the Unilac linear accelerator as an injector. T h e beam intensities are typically between 5.10 4 to 5.10 s particles per spill (with an extraction length of a b o u t 2 s). O n this basis, average rates of a b o u t 100 events are t a k e n for instance for a reaction like Au (150 A MeV) + Au. The target thickness is dictated by two opposing factors. It should not be thin enough to minimize the energy loss for the slower and heavier fragments, but thick enough to provide a reaction to b a c k g r o u n d ratio as high as possible. Typically, one uses Au targets f r o m 0.25% interaction length (96 mg/cm 2 ) to 1% for projectile energies in the 100 to 400 A M e V range. T h e b a c k g r o u n d results essentially f r o m reactions occurring between the b e a m and materials (like H e and C) along its path. O n e percent interaction rate results f r o m the b a c k g r o u n d material u p s t r e a m of the target. T h e air between the target a n d the detector participates with a b o u t a 10% interaction probability. Being mostly light nuclei, they induce low charged-particle multiplicity reactions which stay less t h a n 30 at most. F o r the Au (150 A MeV) + Au reaction the multiplicity extends typically u p to 50 and gets higher for m o r e energic projectiles. T h u s low impact p a r a m e t e r reactions are not appreciably affected by this source of background. O n e millimetre d o w n s t r e a m from the target, b e a m ions a n d reaction p r o d u c t s emitted within a cone of + 30° enter a H e bag (see Figs. 1 a n d 2) m a d e of 100 μπι layered material (polyethylene, aluminium and Mylar) through a 30 μπι K a p t o n window. As already mentioned, this H e bag reduces not only the ion energy loss over the target-detector trajectory b u t also the b a c k g r o u n d due to reactions on all materials but the target.

C. The plastic scintillator wall: 1. The Inner

Wall

It is m a d e of 252 plastic scintillators of 2 cm thickness, trapezoidally shaped a n d read out by 2" p h o t o t u b e s t h r o u g h plexiglass light guides. They are visible in the left part of Fig. 3. T h e exact θ angular coverage of the ensemble is from 1.2° u p to 7.5°. In this way, there is a slight overlap between the Inner Wall and the O u t e r Wall which allows complementary a d j u s t m e n t s of the two submodules f r o m the particles which p u n c h t h r o u g h b o t h detectors. T h e light of a pulsed laser is injected into the center of the rear surface of each scintillator by means of a q u a r t z fiber, for stability checking and calibration purposes as well as a d j u s t m e n t for matching all the modules. The scintillators and light guides are successively w r a p p e d in teflon ribbon (50 μπι), aluminium foil (12 μπι) a n d black tape (100 μπι) to m a k e them light-tight. In provision of P h a s e II, the p h o t o t u b e s have been shielded against a magnetic field of 300 G to keep their functioning u n p e r t u r b e d by any possible fringing field from the superconducting magnet. This is achieved by placing the p h o t o t u b e s inside a m u m e t a l cylinder of 1 m m wall thickness which p r o t u b e s 3 cm past the front glass window of the tube.

Fragment Multi-Detectors

319

Fig. 3. Detail (left) and cutaway assembled (right) views of the mechanical arrangement of the scintillator (Sc), light-guide (LG) and phototube (PT) ensemble of the Inner Wall.

The scintillator-light-guide-phototube elements are placed on 7 concentric rings centered on the b e a m axis according to a mechanical a r r a n g e m e n t illustrated in Fig. 3, in the right part for the general assembly presentation a n d in the left part for a m o r e explicit display. T a b l e 1 summarises the angular coverage of the different rings and paddles and gives the angular binning in polar angle a n d azimuth. O n e sees that the length of the scintillator depends on the ring they c o r r e s p o n d to. F o r design and mechanical reasons each ring is divided into two halves which can be slid up and d o w n along rods as visible in Fig. 3. This allows easy access to the detectors if necessary and permits clearing the b e a m line for experiments at d o w n s t r e a m target stations. The photomultipliers are oriented perpendicularly to the b e a m axis in order to avoid Cerenkov effects and direct particle hits on the p h o t o c a t h o d e s . T h e coordinates of a particle detected by the Inner Wall are determined, for Θ, f r o m the ring of the scintillator p u n c h e d t h r o u g h by the particle, a n d for φ, by the position of

Table 1. Polar 0 and Azimuthal φ Angles (expressed in degrees) subtended by the 7 rings and the paddles of each ring of the inner wall, respectively Ring

1

2

3

4

5

6

7

0mi„ ΑΘ Αφ

1.2 0.78 20

2.0 0.7 15

2.7 0.8 12

3.5 0.8 10

4.3 0.84 8.57

5.1 1.14 7.5

6.2 1.25 6.67

320

Jean-Pierre Coffin

the scintillator in the ring in question. This yields an accuracy in the θ determination of ΑΘ = 0.85° but for φ it is variable depending on the ring. It is equal to 20° for the innermost one (see Table 1). The phototube signals are split into an energy signal and a time signal, delivering directly AE and i slop information. Both the resolution in Δ £ and i stop suffer variations over the surfaces of the scintillators [22-24], In the general case where long and thin parallelepipedic scintillators are used, the amplitude of the signal corresponding to a given excitation decreases rapidly and fairly regularly as the impact gets further away from the read-out end of the scintillator. This is essentially due to increasing loss of light. In the present case also, the trapezoidal shape of the scintillators coupled to short fish-tail light guides creates diffusion processes and multi-reflections which do not favor uniform light collection and yield a response whose amplitude depends on the impact position. This also spoils the AE and f s t o p resolutions. This can be reduced with a rather empirical technique consisting of roughening the scintillator's front face by sandblasting and by replacing the teflon ribbon with black paper. In this way a response fairly independent of the position is obtained. This treatment turned out to be necessary for only the three innermost rings of the Inner Wall. F o r laser pulses of 1 M I L (Minimum Ionizing Light, i.e. the light intensity produced by a particle at its minimum of ionization, it has been verified to remain unchanged irrespective of whether the scintillator light pulses are generated by particles or laser pulses). All paddles of the Inner Wall are found within the range of σ(ί5,ορ) = 180-280 ps. These values decrease only slightly with increasing laser intensity (about 30% between 1 and 400 MIL), indicating that photon statistics at 1 M I L are already sufficient and are not the limiting factor for the time resolution. This problem will be discussed further in Sect. III. C. 2. The overall time resolution is determined by the transition time spread of the light from the different detectors. F r o m cosmic-ray tests an effective light propagation velocity of 13.5 ± 1.5 cm/ns was found which induces a σ(τ) ranging from 230 to 340 ps depending on the length of the scintillator. From the impact coordinates of the detected particle in the scintillator it is possible to calculate its distance to the target which divided by τ gives the particle velocity. This is, however, only the "apparent" velocity since it is affected by some slowing down in the air and all other materials traversed on the way from the target to the detectors. Off-line analysis allows the correction for these energy losses and the exact determination of v. An example of a (Δ£, ν) plot obtained after matching and adding up the 38 paddles of the ring corresponding to = 4.5° for the Au (150 A MeV) + Au reaction is shown in Fig. 4b.

2. The Outer Wall It consists of 512 parallelepipedic strips of scintillator, 18 mm thick by 24 mm high. Each strip is read out at each extremity through a light guide of identical cross-section by 1" photomultiplier. As for the Inner Wall, the scintillators and the light guides are wrapped with Al foil and black tape, and the phototubes are magnetically shielded. The light of the same laser as for the Inner Wall can be injected through a quartz fiber at the center, along the longitudinal dimension, of each strip. The strips are mounted in

Fragment Multi-Detectors

1 a)

1

A u ( 1 5 0 MeV / I I 1 I

u ) +

321

Au Γ

b)

5

5 u

k

Z=1

3 Z=2

. . . 1 . . . 1 "'"«f»mφ»·».., c)

6

..

5 U

\ Vv:.' ν %\ Z= 2 \ \

—10

•

8 ι . I ... I

i a B k

0

8

16

24 0 v (cm /

8 ns)

Fig. 4. Examples of energy loss (ΔΕ) versus velocity (y) plots obtained for Au (150 A MeV) + Au with the Rosace (a), the Inner Wall (b), the Parabola (c), and the Outer Wall (d).

8 separated sectors of 64 elements as shown in a perspective view presented in the left part of Fig. 5. F o r a sector the strips are positioned in such a way that the light guides and tubes of one sector are hidden by active areas of the two n e i g h b o u r i n g sectors. All strips are oriented so that their entrance face is o r t h o g o n a l in its center to the target-strip axis. These features yield a n o n - p l a n a r geometry with zigzag borderlines between the sectors as shown in the right p a r t of Fig. 5. T h e active scintillator length varies f r o m 45 cm to 165 cm for Τ —

Τ

i°" , ,!,>>>>>)

0

!)> >\> >n η ,

0.5

1 V/

Yp

Fig. 17. Acceptance and thresholds of the 4n-Detector (Phase I) expressed in the transverse m o m e n t u m per nucleon pJA (normalized to projectile m o m e n t u m per nucleon p p /A p ) versus rapidity plot, scaled with respect to projectile rapidity y p . T h e hatched regions (right slanted) show the inefficient zones of detection: θ < 1°, θ > 30°, and a r o u n d 7° and 19°, i.e. in the shadow created by the structure supporting the Parabola. T h e hatched sectors (left slanted) represent the domains where the rapidity of Ζ = 3 ions is below the experimental threshold, as calculated for the Au (150 A MeV) + Au reaction. Thresholds for Ζ = 1 and 2, and Ζ = 8 are also indicated with thick solid and dashed curves, respectively.

Table 2. Lower energy thresholds (in MeV/nucleon) for the detection of various Ζ fragments for the rosace-inner wall and the parabola-outer wall couplings Ζ

1

2

3

4

5

6

9

12

Rosace-Inner Wall thresholds (A MeV) P a r a b o l a - O u t e r Wall thresholds (A MeV)

21

21

26

31

38

42

52

62

14.5

14.5

17

20

24

28

33

40

the lab system rises. The lower energy thresholds (in MeV/nucleon) for the RosaceInner Wall and Parabola-Outer Wall couplings are presented in Table 2. They are given for the mean angle of the angular domain covered by these two detector ensembles. These thresholds have been calculated with an energy-loss code which is based on the Bethe theory higher order corrections [34], taking into account all materials such as air, He, the different foils traversed and half the target thickness. This energy-loss code is also used in the final data analysis to correct the "apparent" velocities of the measured products. The closer to the threshold and the heavier the fragment, the more important and necessary the correction.

Fragment Multi-Detectors

339

VII. Background As far as background is concerned, one only has to worry about unwanted or incorrect information within a valid event, i.e. after a trigger on a specific event type has been set. Reactions of the beam on material other than the target have already been discussed in Sect. III. B. 2. They are negligible for high-multiplicity events. The mixing of fragments produced by different beam particles is excluded by a "double beam" bit. It is set each time that two t l signals (cf. Sect. III. B. 2) are registered within a specific time normally taken as 10 μβ. Thus too high a beam intensity will increase the average load on the detectors but will not lead to the registration of intruder fragments. There remain essentially two sources of unwanted hits within an event: 1) The fragments can interact on their way to the detectors or in the detector material itself, the second case being the more probable. The reaction rate with the detector material, carbon nuclei mainly, increases from less than 1 % for protons to about 10% for Ζ = 8 fragments. If the reaction occurs in the periphery of the detector, the incoming fragment may be misidentified. This depends on how much the total scintillation light intensity of all secondary products differs from that produced by the primary fragment. It depends also on whether the secondary reaction products leave the detector without hitting their neighbours. Central impacts, which represent however only a small fraction of the total reaction rate in the detector material, will trigger nearby detectors in many cases. This background rate cannot be mastered and has to be tolerated. 2) There are hits registered in the detectors by neutrons and particularly 1 4 0 * ( l ~ ) - > 1 3 N + p reaction was observed by detecting 1 3 N and protons in their final state [44,45]. Figure 19 shows an example of the measured angular distribution of the reaction. Although the 1 3 N(p, y ) 1 4 0 reaction energy of interest is low, the best beam energy for this inverse reaction is ~ 90 A MeV. Therefore one can use a thick target (350 mg in ref. 44) to increase the detection yield drastically. Moreover the statistical factor of the final state favors the inverse reaction. Results of the capture width of the 1 3 N(p, y ) 1 4 0 reaction are summarized in Fig. 20. The results obtained by the different methods agree well with each other. F o r rapid neutron capture processes (n, y) reactions play an important role. It is extremely difficult, if not impossible, to observe neutron scattering of a β unstable nuclei. However inverse reactions with radioactive beams (virtual y, n), the same technique shown above, could possibly be studied.

2. Big Bang

nucleosynthesis

Although standard Big Bang nucleosynthesis involves virtually no reaction of shortlived nuclei, recent investigations have forced on nonstandard models, one class of

Production and Use of Radioactive Beams

367

®cm Fig. 19. Angular distribution of γ capture process determined from 1 4 0 + Pb-> 1 3 N + ρ + χ reaction. The same measurement was made for 1 3 N + Pb —•12C + ρ + χ reaction. Agreement of the measured capture width with the inverse reaction 1 2 C + p - > 1 3 N + y which had been measured in good accuracy, provided a verification of the method.

which is the i n h o m o g e n o u s model [ 4 6 , 4 7 ] . In this model, p r o t o n rich regions a n d neutron rich regions are produced a few m o m e n t s after the q u a r k - h a d r o n p h a s e transition. Reaction n e t w o r k calculation for the n e u t r o n rich region suggests that all the nuclides, a l t h o u g h they are in small a m o u n t s c o m p a r e d with the solar a b u n d a n c e , of the periodic table would be synthesized in the first few minutes of the universe. Figure 21 shows a n e t w o r k for synthesis of some of the light nuclides. It shows that several reactions involving 8 Li are important. All nuclides heavier t h a n A = 11 are funneled t h r o u g h " B o n their nucleosynthesis paths a n d 1 J B is formed p r e d o m i n a n t l y via the 8 L i + 4 H e - > n B + η reaction. T h u s this reaction rate as well as the reaction rates to destroy 8 Li are crucial to making accurate predictions with i n h o m o g e n e o u s models. T h e cross section of the 8 Li + 4 H e - > n B + η reaction was first determined by Paradeiiis et al. [48] using the inverse reaction J B + η - » 8 L i + 4 He). H o w e v e r the

368

Isao T a n i h a t a

a

b

o

d

e

Fig. 20. G a m m a width of 1 3 N(p, γ ) 1 4 0 , reaction measured by various methods, a and b are indirect determinations by the measurements of the branching ratio, c and d are from the inverse reaction method, and e is the direct measurement. They agree well within error bars.

Fig. 21. Partial reaction network for Big Bang nucleosynthesis of Ζ = 3 - 5 nuclides. Reactions that either produce or destroy 8 Li are crucial because 8 Li appears to be central to the formation of heavier nuclides in the inhomogeneous models.

inverse reaction method restricts the measurement of the transition only to the ground state of n B . On the other hand the reaction rate of interest includes all the transitions which go to the excited state o f 1 1 Β and reach the ground state. A direct measurement of this reaction using a beam of 8 Li was recently reported [49]. Figure 22 shows the determined cross section. It was found that the cross sections are about four to five

Production and Use of Radioactive Beams

369

+ »Li (α ,η o 8Li (a,n)11B(g.s: 600

|400

Μ

200

O o o o 0 ° 0 ° 0 o o o C b ,Όο 0 0

2

4

ο OOP 6

E c m (MeV) Fig. 22. Cross section of 8 Li(a, n)1 ' B reaction. Open circles show the one obtained by the inverse reaction 1 'B(n, y) 8 Li that include the transition only to the 1 'B g s. Reaction cross section is about 4 - 5 times larger than those of ground state transition.

times larger than that of ground state transition. Also other related reactions [ 8 Li(d, t) 7 Li, 8 Li(d, p) 9 Li, 8 Li(d, n) 9 Be] were recently measured [50]. The results of the recent calculation of the primordial abundance using these data are shown in Fig. 23 [51]. Considerable enhancements of the atomic abundance are seen for elements heavier than boron.

C. Radioactive beams for other fields of study Low energy radioactive beams have been used for many studies of different disciplines. The high-energy radioactive beams are new and have not yet been used in other fields of research than nuclear physics, astrophysics and cancer therapy [52], However they are considered to be extremely powerful tools for studies in many disciplines. Here, I show the important properties of radioactive nuclei beams for possible applications below: 1. Elemental selection: Any element may be delivered independent of its chemical properties and therefore the best element can be used for a specific problem. A wide variety of elements is possible for appropriate uses. For example, interstitial sites of crystals can be studied selectively by many available elements. 2. High energy: Because the energy is high, the beam is easily controlled by magnetic system. Also implantation depth can be adjusted by changing the beam energy. Therefore, it provides for example, the first opportunity to start a tracer from any depth independent of an elemental composition of the sample material.

370

Isao Tanihata

Mass Number A Fig. 23. Result of model calculations of atomic abundance. Inhomogeneous model predict more production of heavier elements.

3. Radioactivity: By definition, a radioactive ion decays in the material where it stopped. Emissions of α, β, and/or y rays can be used for probing the stopped position and the surrounding electric and magnetic fields with extremely high sensitivity. 4. Wide range of lifetime and spin: A selection of a lifetime enables the application to phenomena of various time range (1 μs-years). A selection of nuclear spin is also possible within an element. 5. Polarized beam: Polarized beams of radioactive nuclei can also be produced by studying the magnetic and electric structure of the material. Probes which have some of the advantages listed above existed before but none of them had them all. The R N B has all of these advantages and large possibilities of application.

References [1] Radioactive Beams, Proc. of the First Int. Conf. on Radioactive Nuclear. Beams, Berkeley, Ca 1989, Myers, W. D„ Nitschke, J. M., Norman, Ε. B. Eds., World Scientific, 1990. Proc. of the second Int. Conf. on Raidoactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar,

Production and Use of Radioactive Beams

[2]

[3]

[4]

[5] [6] [7] [8] [9]

[10] [11] [12]

[13] [14] [15]

[16] [17]

[18]

[19] [20] [21]

371

Th., Eds., Adam Hilger 1992. Proc. of the Workshop on the Science of Intense Radioactive Ion Beams, Los Alamos, 1990, LA-11964-C UC-413. Proc. of Int. Symp. on Structure and Reactions of Unstable Nuclei, Niigata, Japan 1991, Ikeda, K. and Suzuki, Y., Eds., World Scientific. Proc. of the Int. Workshop on Unstable Nuclei in Astrophysics, Tokio, Japan 1991, Kubono, S. and Kajino, T., Eds., World Scientific. Nucleus-Nucleus Collision IV, Toki, H., Tanihata, I., and Kamitsubo, H., Eds., Special Issue of Nuclear Physics, North-Holland, Nucl. Phys. A538 1992. Tanihata, I., On the Possible Use of Secondary Radioactive Beams, Treatise on Heavy-Ion Science, Vol. 8, Bromley, D. Α., Eds. (Plenum, New York 1989) p. 443. Detraz, C., Vierira, D. J., Exotic Light Nuclei, Ann. Rev. of Nuclear and Particle Science (1989) p407. Heckman, Η. H., Greiner, D. E., Lindstrom, P. J., Shwe, H., Fragmentation of 4 He, 12 C , 1 4 N , and 1 6 0 Nuclei in Nuclear Emulsion at 2.1 GeV/nucleon, Phys. Rev. C 17 (1978) 1735. Greiner, D. E., Lindstrom, P. J., Heckmann, Η. Η., Cork, Β., Bieser, F. S., Momentum Distributions of Isotopes Produced by Relativistic 1 2 C and l e O projectiles, Phys. Rev. Lett. 35 (1975) 152. Goldhaber, A. S., Heckmann, Η. Η., High Energy Interactions of Nuclei, Ann. Rev. Nucl. Sei. 28(1978) 161. Gelbke, C. K., et al., Similarity of Cross Sections for Peripheral Collisions at 20 MeV/A and 2.1 GeV/A, Phys. Rev. Lett. 37 (1976) 1191. Van Biber, Κ. et al., Evidence for Orbital Dispersion in the Fragmentation o f 1 6 0 at 90 and 120 MeV/nucleon, Phys. Rev. Lett. 43 (1979) 840. Alonso, J. R., Chatterjee, Α., Tobias, C. Α., IEEE Trans. Nucl. Sei. 26 (1979) 3003. Tanihata, I., Nuclear Physics Using Unstable Beam, Hyperfine Interactions, 21 (1985) 251. Tanihata, I. et al., Measurement of Interaction Cross Sections and Radii of He Isotopes, Phys. Lett. 160B (1985) 380. Dufour, J. P. et al., Projectile Fragment Separation: Application to the LISE Spectrometer at GANIL, Nucl. Instr. Meth. in Phys. Res. A56 (1986) 34. Geissei, H. et al., The GSI Projectile Fragment Separator (FRS): a versatile magnetic system for relativistic heavy ions, Nucl. Instr. Meth. in Phys. Res. B70 (1992) 286. Müller, A.C., Anne, R., LISE 3: A magnetic spectrometer - Wien filter combination for secondary radioactive beam production, Nucl. Instr. Meth. in Phys. Res. B70 (1992) 276. Kubo, T. et al., The RIKEN radioactive beam facility, Nucl. Instr. Meth. in Phys. Res. B70 (1992) 309. Asahi, K. et al., Hyperfine Interaction in press 1992. Smith, R. J., et al., Production and Use of Radioactive 7 Be Beams, Phys. Rev. C43 (1991) 761. Smith, R. J., et a l . Scattering of 6 He from 1 9 7 Au, na, Ti, 27A1, na, C, and 9 Be at Ε = 8 - 9 MeV, Nucl. Instr. Meth. A294 (1990) 26. Sawicki, J. A. et a l . The Isospin Laboratory proposal, LALP 91-51,1991. Darquennes, D. et a l . Production of intense radioactive ion beams using two accelerators, Phys. Rev. C42 (1990) R804. Darquennes, D. et al, A high temperature graphite target for the production of 1 3 K, Nucl. Instr. Meth. Res. B47 (1990) 311. Decrock, P. et a l . On electron cyclotron resonance ion source for efficient production of radioactive ion beams, Nucl. Instr. Meth. Phys. Res. B58 (1991) 252. Decrock, P. et a l . Production and acceleration of radioactive ion beams at energies of astrophysical interest with the Louvain-laNeuve cyclotrons, Proc. of the Second Int. Conf. on Radioactive Beams, Louvian-laNeuve, Belgium, August, 1991, Adam Hilger Delbar, Th. E d , p. 121. Ravn, H. L , Radioactive nuclear beam facilities on ISOL-post accelerator schemes, Proc. of the Second Int. Conf. on Radioactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar, Th. E d , Adam Hilger 1991, p. 85. Takigawa N , Sagawa, H , Interaction potential and fusion of a halo nucleus, Phys. Lett. Β 265 (1991) 23. Kobayashi, T , Proc. of Int. Symp. on Structure and Reactions of Unstable Nuclei, Niigata, Japan 1991. Ikeda, K„ Suzuki, Y„ Eds, World Scientific, pl87. Tanihata, I. et a l . Measurement of Interaction Cross Sections and Radii of p-Shell Nuclei,

372

[22] [23] [24] [25] [26] [27]

[28]

[29]

[30]

[31] [32] [33] [34] [35] [36] [37]

[38]

[39]

[40]

[41] [42]

[43]

Isao T a n i h a t a Phys. Rev. Lett. 55 (1985) 2676. Tanihata, I., Structure of N e u t r o n Rich Nuclei Studied by Radioactive Beams, Nucl. Phys. A522 (1991) 275c. Kobayashi, T. et al., Projectile Fragmentation of the Extremely N e u t r o n Rich N u c l e i 1 ' L i at 0.79 GeV/nucleon, Phys. Rev. Lett 60 (1988) 2599. Hansen P. G., Johnson, B., The N e u t r o n Halo of Extremely Neutron-Rich Nuclei, Europhysics Lett. 4 (1987) 409. F u k u d a , M. et al., N e u t r o n Halo in 1 'Be Studied Via Reaction Cross Sections, Phys. Lett. Β 268 (1991) 339. Tanihata, I. et al., Determination of the Density Distribution and the Correlation of Halo N e u t r o n s in 1 'Li, Phys. Lett. Β 287 (1992) 307. Shimoura, S., Proc. of Int. Symp. on Structure and Reactions of Unstable Nuclei, Niigata, J a p a n 1991, Ikeda, K. and Suzuki, Y., Eds., World Scientific, p. 132. Kobayashi, T. et al., Electromagnetic Dissociation and Soft Giant Dipole Resonance on N e u t r o n Dripline N u c l e i 1 1 Li, Phys. Lett. Β 232 (1989) 51. Kobayashi, T., Proc. of the First Int. Conf. on Radioaxtive Nuclear Beams, Myers, W. D., Nitschke, J. M. and N o r m a n , Ε. B., Eds., World Scientific Pub. 1990, p. 325. Warner, D. D., Nuclear Physics with Radioactive Beams, Proc of the Second Int Conf. on Radioactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar, Th., Ed., A d a m Hilger 1991 p. 139. Ikeda, K., Structure of Neutron Rich Nuclei, Nucleus-Nuclues Collision IV, Toki, H., Tanihata, I., Kamitsubo, H., Eds., Special Issue of Nuclear Physcis, North-Holland, Nucl. Phys. A538 (1992) 355c. Esbensen, H., Correlated Dipole Response of N e u t r o n Rich Nuclei, Proc. of Int. Symp. on Structure and Reactions of Unstable Nuclei, Niigata, J a p a n 1991, Ikeda, K. and Suzuki, Y., Eds., World Scientific p. 178 and references therein. Hoshino, T. et al., G i a n t Resonances of Light N e u t r o n Rich Nuclei, Nucl. Phys. A523 (1991)228. Tosaka, Y., Suzuki, Y., Structure of " L i in the cluster orbital shell model for the 9 Li + η + η system, Nucl. Phys. A512 (1990) 46. Beiner, M., L o m b a r d , R., Mas, D., Self-Consistent Calculations of G r o u n d State Properties for Unstable Nuclei, Nucl. Phys. A249 (1975) 1. Hirata, D. et al., Relativistic Hartree Theory for Nuclei far from Stability Line, Phys. Rev. C44 (1991) 1467. Tanihata, I. et al., Revelation of Thick N e u t r o n Skin, Phys. Lett. B289 (1992) 261. M o o n , C. B. et al., Measurement of 11 Li + ρ and 9 Li + ρ Ellastic Scattering at 60 MeV, Phys. Lett. Β in press. Baye, D., Measurement of t h e 1 3 N + 1 2 1 3 C elastic scattering, Proc. of the Seocnd Int. Conf. on Radioactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar, Th. Ed., Adam Hilger 1992, p. 173. Terenetski, K. O. et al., 86 MeV 9 Li elastic scattering by 2 0 8 P b , Proc. of the Second Int. conf. on Radioactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar, Th. Ed., Adam Hilger 1992, p. 231. Kolata, J. J., Experiments with radioactive nuclear beams at the University of Notre Dame, Proc. of Int. Symp. on Structure and Reactions of Unstable Nuclei, Niigata, Japan 1991, Ikeda, K. and Suzuki, Y., Eds., World scientific, p. 252. Tanihata, I., Kobayashi, T. Yamakawa, O., Shimoura, S., Ekuni, K., Sugimoto, K., Takahashi, N., Shimoda, T., Sato, H. Measurement of Interaction Cross Sections using Isotope Beams of Be and Β and Isospin Dependence of the Nuclear Radii, Phys. Lett. B206 (1988) 592. Sato, H., O k u h a r a , Y., N u c l e u s - N u c l e u s Scattering and Interaction Radii of Stable and Unstable Nuclei, Phys. Rev. C34 (1986) 2127. Barnes, C. Α., Nuclear astrophysics with radioactive nuclear beams, Proc. of the Second Int. Conf. on Radioactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar, Th. Ed., Adam Hilger 1992, p251 and references therein. Decrock, P. et al. Measurement of the ' Η ^ ' Ν ^ δ ^ Ο Cross Section and Determination of

Production and Use of Radioactive Beams

[44] [45]

[46]

[47] [48] [49] [50]

[51] [52]

373

14 0 ( 1 ~ , Τ— 1, E exc = 5.17 MeV) γ-width, Proc. of the Int. Workshop on Unstable Nuclei in Astrophysics, Tokio, Japan 1991, Kubono, S., Kajino, T., Eds., World Scientific, p. 75. Benjelloun, M. et al., Precise measurement of 1~ resonance state of 1 4 N by the 1 3 C + ρ reaction, Proc. of the Second Int. Conf. on Radioactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar, Th. Ed., Adam Hilger 1992, p265 and references therein. Decrock, P. et al., Determination of the 13 N(p, y ) 1 4 0 Reaction Cross Section using a 1 3 N Radioactive Ion Beam, Phys. Rev. Lett. 67 (1991) 808. Motobayashi, T. et al., Determination of the astrophysical 13 N(p, γ ) 1 4 0 cross section through the Coulomb dissociation method, Phys. Lett. B264 (1991) 259. Kiener, J., Determination of the 13 N(p, γ) reaction rate through Coulomb break-up of a radioactive beam, proc. of the Second Int. Conf. on Radioactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar, Th. Ed., Adam Hilger 1992, p. 311. Applegate, J. H., Hogan, C. J., Scerrer, R. J., Cosmological Quantum Chromodynamics, neutron diffusion, and the production of primordial heavy elements, Astrophysics J. 329 (1988) 592. Mathews, G. J., Meyer, B. M„ Alcock, C„ Fuller, G. M„ Coupled Baryon Diffusion and Nucleosynthesis in the early Universe, Astrophys J. 358 (1990) 36. Kajino, T., Boyd, R. N., Production of the light elements in primordial nucleosynthesis, Astrophys. J. 359(1990) 267. Paradeiiis, T. et al., Astrophysical S(E) factor of 8 Li(a, No)11 Β and inhomogeneous Big Bang nucleosynthesis, Z. Phys. A337 (1990) 211. Boyd, R . N . et al., Measurement of 8 Li(a, η ) η Β reaction cross section at energies of astrophysical interest, Phys. Rev. Lett. 68 (1992) 1283. Farrel, Μ. M., et al., 2 H induced reactions on 8 Li and primordial nucleosynthesis, Proc. of the Second Int. Conf. on Radioactive Nuclear Beams, Louvain-la-Neuve 1991, Delbar, Th. Ed., Adam Hilger 1992, p. 287. Kajino, T., private communication. Chatterjee, A. and Llacer, J., Proc. of the First Int. Conf. on Radioactive Nuclear Beams, Berkeley, Ca 1989, Myers, W. D., Nitschke, J. M., Norman, Ε. B. Eds., World Scientific 1990, p. 403.

11 In-Flight Separation of Heavy Ion Beams Gottfried

Münzenberg

Table of Contents I. Introduction 375 A. Sources of energetic secondary beams 377 B. Reaction kinematics 378 C. Luminosity 382 II. Separation techniques 383 A. Ion optics for in-flight separators 384 B. Liouvillean ion optics 387 C. Design principles 389 III. Examples of in-flight separators 391 A. Recoil mass separators 391 B. Velocity filters 396 C. RF separators 399 D. Parabola spectrographs 401 E. Gasfilled separators 402 IV. Separation of relativistic heavy ions 404 A. The energy loss separators for projectile fragments 404 B. The energy degrader as an ion-optical element 408 C. Separators for projectile fragments 411 D. Secondary beam facilities 413 E. Cooling of in-flight separated secondary beams 418 V. Conclusion 420 References 420

I. Introduction In this review the separation techniques for ions produced in nuclear reactions will be discussed with the restriction to in-flight separation where the reaction products pass an electromagnetic separator almost unretarded with their kinetic energy from the nuclear reaction process leading to their creation. Consequently these separators are kinematic separators - part of them even uses the reaction kinematics to select the desired nuclear species - hence the quality of the separated particle beams, such as kinetic energy or phasespace population are determined by the kinematics of the production process. Recently storage rings with beam cooling are used to improve the phasespace density and to allow for de- or acceleration of the stored beams of in-flight separated heavy ions, which for the first time allows the preparation of in-flight separated beams with a quality independent of their reaction kinematics.

376

Gottfried Münzenberg

Fig. 1. Halflife domains accessible with in-flight separation compared to on-line separation (Isol), heliumjet gas transport system and chemical methods.

The advantage of in-flight separation is the short separation time, which essentially is determined by the flight time through the separator. It is of the order of milliseconds. Figure 1 shows a comparison of the halflife ranges covered with the various separation techniques for the detection of instable nuclei. The in-flight separation combined with appropriate detector systems covers the largest span of about twelve orders in magnitude. In-flight separated particles keep their full kinetic energy from the nuclear reaction, sufficient for energy loss, time-of-flight or total kinetic energy measurement which permits a high redundancy in their identification and is helpful for the application of delayed-coincidence techniques. Consequently in-flight separation could be developed to ultimate sensitivity, the lowest rate for an isotope to be identified was as low as one atom per week [1], The first recoil separators have been applied to the separation of fission fragments. They were installed at nuclear reactors as for example the Munich separator, a classic M a t t a u c h - H e r z o g spectrometer [2], or various types of gas filled separators [3, 4]. Later, after powerful heavy-ion accelerators went into operation, instruments for the separataion of heavy-ion fusion products were developed either as velocity filters [5] or recoil mass separators [6], The most recent instruments are used for relativistic projectile fragments [7-10]. The separators especially for fission products and recoils from heavy-ion fusion have been reviewed in a number of papers [11,12], fragment separators are described in [9]. This paper will emphasize the recent developments of separators for recoils from heavy-ion fusion and the separation of relativistic nuclei created by projectile fragmentation. After a general and short overview on the production processes and the reaction

In-Flight Separation of Heavy Ion Beams

377

kinematics the various separation techniques and their limitations will be discussed. Together with a brief introcuction into the ion optics and basic design principles some characteristic examples of in-flight separators and their principal applications will be described. Finally an outlook on new techniques applied to in-flight separated particles such as storage and cooling in a storage ring will be given.

A. Sources of energetic secondary beams As already mentioned, the historic source of energetic nuclei to which in-flight separation has been applied was neutron induced fission of heavy targets like uranium placed in high-flux reactors. Accelerator based sources for in-flight separated nuclei are proton induced uranium spallation and complete fusion of heavy-ions, respectively. With the development of accelerators for relativistic heavy-ions, nuclear fragmentation became a new tool for the production of energetic beams of instable heavy nuclei with energies far above the C o u l o m b barrier. Secondary beams of that kind can be directly used for nuclear reactions without further accelearation [9], The regions of the chart of nuclei accessed by the various production methods are shown in Fig. 2. In neutron induced uranium fission a light and a heavy group of nuclides, around krypton and xenon are pronounced. This reaction leads to neutron

Fig. 2. Regions of the chart of nuclei which can be accessed by induced fission, complete fusion of heavy ions, and fragmentation.

378

Gottfried Münzenberg

rich isotopes because of the neutron excess of the fissioning heavy nucleus. Proton induced uranium fission creates symmetric fragments with a peak cross section near the neutron rich palladium isotopes. Heavy-ion fusion, the amalgamation of two nuclei, generally leads to the region of neutron deficient isotopes, as a heavy nucleus in a region of large neutron excess is made from the constituents of two light nuclei. This is the only type of reaction discussed here which leads to products heavier than either of the initial nuclei. Consequently the heaviest known elements have been produced by heavy-ion fusion. The projectile fragmentation is in principle the spallation in inverse kinematics: a heavy relativistic projectile hits a light target. The heavy fragments emerging from peripheral collisions generally are exotic nuclei, so the stable projectile beam is converted into a beam of predominantly instable constituents, isotopes which are lighter than the projectile are produced [10]. The fragmentation reaction is a hot process leading to the evaporation of a great number of neutrons. Therefore the cross sections peak along the neutron deficient side of the chart of nuclides. For the production of relativistic neutron rich nuclei another process recently has been considered, the electromagnetic dissociation of fissile projectiles, which in principle is photodisintegration by the equivalent photons created in the electromagnetic field which a relativistic nucleus experiences when passing by a highly charged target [13]. For relativistic energies and targets with high nuclear charges such as lead the cross sections become of the order of the geometric cross section or even larger. In general for all reactions described above the cross sections and the production rates, respectively, decrease rapidly towards the limits of nuclear stability, e.g. the drip-lines or, in the heavy-element region, when the limits of nuclear fissility are reached. All reactions discussed here lead to broad isotopic distributions, with exception of the complete fusion, where only few isotopes close to the compound nucleus are formed by particle evaporation from the nucleus amalgamated from projectile and target. Sophisticated techniques are needed to achieve a really clean isotopic separation. In addition to the classical mass-separator designs such as the Mattauch Herzog separator or the parabola spectrograph, velocity filters in combination with magnetic spectrometers or time-of-flight separators have been developed. Non-Liouvillean separators, such as gas filled separators or the separators for relativistic ions take advantage of the nuclear-charge dependence of the interaction of fast particles with matter.

B. Reaction kinematics The kinematic properties of the reaction products such as average energy, energy spread, and angular distribution are determined by the mechanism of the nuclear reaction. Except for the projectile fragmentation where the reaction products simply have a velocity close to that of the projectile, the intrinsic nuclear forces, predominantly the Coulomb energy, determine the reaction kinematics. Figure 3 gives an overview over the energy domains of the various production processes discussed here.

In-Flight Separation of Heavy Ion Beams

379

I& § 102 V)

ω (0 Ε 10 1 CT £ 10°

LU 10"

20

40

60

80

100

Element number Fig. 3. Domains of kinetic energies of the reactions products f r o m various nuclear reactions.

Fission fragments gain their energy from the Coulomb repulsion of the fragments in the phase of their separation. The recoil energy of a fragment with mass Au from the mother nucleus with mass A is obtained with the help of the Viola systematics [14]: £ f r = 0 . 1 2 - ^ 0 4 - ^ ) MeV.

(1)

Heavy-ion fusion is the complete amalgamation of target and projectile nuclei, hence the newly created compound nucleus suffers the full momentum transfer from the projectile. It recoils from the target with center of mass velocity. The energy is: £ £r=X P A

(2)

c

with Ap and Ep denoting projectile mass and energy, respectively, Ac standing for the mass of the compound nucleus. For the production of exotic species the projectile energy is chosen as low as possible, e.g. close to the Coulomb barrier to produce the fragile compound nuclei with the lowest possible excitation energy. With a Coulomb barrier near E p x 5 MeV/Λ. the recoil energy is: A2 ET = 5 — MeV.

(3)

380

Gottfried Münzenberg

T h e velocity of projectile fragments is essentially determined by the velocity of the projectile. T h e energy is:

^

=

λ

(4)

Ρ

These considerations prove that all sources of secondary beams have energies below the C o u l o m b barrier (Fig. 3) and c a n n o t be used for secondary reactions without further acceleration except for projectile fragments, which may have energies u p to the GeV/u range. T h e energy of projectile fragments is determined by the energy of the projectiles, e.g. by the accelerator to produce the projectile beam. Cyclotrons typically deliver b e a m s u p t o 0.1 GeV/u, higher energies are available at synchrotrons. The figure also shows the energy where 5 0 % of the ions are b a r e which is an i m p o r t a n t limit for efficient a n d clean ion optical separation. T h e separation efficiency, transmission as well as the purity of the separated beams, depend on the phase-space population of the reaction p r o d u c t s such as angular distribution a n d relative m o m e n t u m spread. Figure 4 c o m p a r e s the acceptances of typical separator types to the typical widths of the particle distributions. Acceptances of spectrographs range u p to several tens of millisteradians and several tens of percent in angle and relative m o m e n t u m spread, respectively. Spectrometers typically accept less t h a n 10 millisteradians and less t h a n a b o u t 5 % in relative m o m e n t u m spread.

solid angle [msr] Fig. 4. Solid angle and momentum spread of the reaction products compared to the acceptances of spectrographs and spectrometers.

In-Flight Separation of Heavy Ion Beams

381

As m o s t of the nuclear reaction products are created at large excitation energies they c o o l d o w n to the groundstate predominantly by e v a p o r a t i o n of n u c l e o n s and alphas. T h e widths of the energy distributions are determined by the recoil from these particles. Fission products are an exception, they are emitted into the full solid angle and their energy spread in the laboratory frame is essentially determined by the total kinetic energy of the fragments. Similar considerations hold for spallation products. T h e efficiency for the in-flight separated fission products is as small as 1 0 " 6 [ 1 2 ] . Evaporation residues from heavy-ion fusion are kicked out of the target by the m o m e n t u m transfer they suffer from the projectile. T h e y are emitted into a narrow c o n e in b e a m direction and with a small velocity spread. T h e phase-space p o p u l a t i o n of the recoils is determined by the ratio of the sum of the recoil m o m e n t a from the particles evaporated from the hot c o m p o u n d nucleus related to the projectile m o m e n t u m . W i t h increasing m o m e n t u m e.g. mass of the projectile the kinematic focusing of the evaporation residues increases. If a number of η particles is evaporated, the relative m o m e n tum spread of the recoils is [ 5 ] :

- = — Pr Ρρ

(5)

where p p and p t are mass and m o m e n t u m of the projectile and of the i-th e v a p o r a t e d particle, respectively. Typical for fusion-evaporation products created with sufficiently heavy beams (A > 40) and with b e a m energies close to the C o u l o m b barrier are solid angles of less than 5 msr and velocity spreads of few percent in the case of neutron evaporation only. Consequently the separation efficiencies are 10% to 100%. T h e m o m e n t u m distribution of projectile fragments reflects the Fermi m o m e n t a of the nucleons abraded in the nuclear collision and the m o m e n t a of the e v a p o r a t e d nucleons. T h e relative m o m e n t u m spread in b e a m direction can be obtained with the systematics from Morrissey which nicely reproduces the m o m e n t u m distributions of relativistic projectile fragments. T h e longitudinal m o m e n t u m spread is: σ

ff

=

o\A4p-/lfr,

(6)

with the empirical constant σ0 = 87 M e V / c . E q u a t i o n (6) s h o w s that the m o m e n t u m spread of projectile fragments is independent of the projectile energy, hence the kinematic focusing of projectile fragments increases with the projectile energy. F r o m Eq. 6 we obtain with p f r = ßfr yfr E0/c :

Pfr

Pfr

E0 is the rest energy of the fragment of 931 A{l GeVw, ß(T = vfT/c, and yu = l / ^ / l — ßfr. T h e relative m o m e n t u m spread grows with the increase in the n u m b e r of abraded nucleons and for light projectiles. T h e m o s t dense phase-space p o p u l a t i o n s are

382

Gottfried Münzenberg

obtained from fragments close to the projectile and in the region of heavy nuclei. Typical values for solid angle and relative m o m e n t u m spread for heavy and relativistic fragments with A > 100 and E/u = 500 MeV/w are 1 msr and 2 % , respectively. This fits well into the acceptance of high resolving spectrometers (Fig. 4). F o r projectile fragments in the energy regime above 100 MeV/u separation efficiencies of 50% to 100% have been achieved.

C. Luminosity The sensitivity of a set-up is determined by the source strength and the separator transmission. The source strength can be described in terms of the luminosity L= Nv JV t

(8)

where N p and AT, are the beam particle current on the target and the target thickness in atoms per square centimeters, respectively. The production rates are calculated from the product of luminosity and cross section. The secondary beam intensity is determined by the neutron flux in the nuclear reactor or the beam current of the accelerator, respectively. High flux reactors provide neutrons with a particle density of 5.10 1 4 /cm 2 . The beam currents at modern heavy-ion accelerators of the linac or cyclotron type exceed 10 13 /s, synchrotron beams have by 3 to 4 orders of magnitude lower intensities due to the small duty factor of these periodically accelerating machines. New, pulsed high-current sources are under development to increase synchrotron-beam intensities. Using in-flight separation the target thickness is generally limited, as the energy loss and the straggling of projectile beam and reaction products due to atomic interactions with the target material spoils the kinematic properties such as energy and angular spread, respectively [15]. The commonly used target thicknesses for the production of fission fragments and heavy-ion fusion products, respectively, are of the order of mg/cm 2 which corresponds to about 10 18 atoms per square centimeter. Projectile fragmentation profits from the high energy of beam and fragments. Depending on the projectile energy, typical target thicknesses range from O.lg/cm 2 to 10g/cm 2 , corresponding to 10 21 to 10 2 3 atoms per square centimeter. The geometric nuclear cross sections being of the order of barns, the thickness is limited by self-absorption of the reaction products in the target. Consequently the luminosities range from about 10 3 7 ( c m 2 s ) for fission and fusion, respectively and up to the order of 10 3 6 / (cm 2 s) for projectile fragments. The lower limits of cross sections to be detected with a sensitive separator of a moderate transmission of 10% are 10 pb for one atom per day for heavy ion fusion and 10 pb for one atom per second for fragmentation, respectively. The production cross sections for a specific nuclear species not too far from stability range from 10 m b to 1000 mb, d r o p to the picobarn region for the most exotic species at the drip-lines or in the heavy-element region. Table 1 summarizes the estimated particle rates for the production methods and related separation techniques. Fission sources, though profiting from the high neutron flux and the high fission cross sections, suffer from the small transmission of the recoil separators, consequently only rates up to 10 5 /s

In-Flight Separation of Heavy Ion Beams

383

Table 1. Secondary beam intensities from various nuclear reactions

Projectiles on target ( s ~ ' ) Target a t o m s (cm ~ 2 ) Luminosity (s " 1 cm ~ 2 ) Production cross-section (cm 2 ) Sec. rates ( s ~ ' ) Separation efficiency Intensity of separated beam

Ind. Fission

HI-fussion

Projectile F r a g m e n t a t i o n E/w < 100 MeV/u E/u > 100 MeV/u

51014 10 1 8 5-10 3 2 < 10"23 < 5109 10 4 — 10 < 510s

10 1 3 10'8 10 3 1