EXOGAM is a European collaboration to build a highly efficient and powerful g-ray spectrometer for nuclear spectroscopy using the exotic radioactive beams from the Systeme de Production d'Ions Radioactifs et d'Acceleration en Ligne (SPIRAL) facility at the Grand Accelerateur National d'Ions Lourds (GANIL).
This Annex gives a description of the EXOGAM project which is a high efficiency array of g-ray detectors with associated high voltage, autofill system, cabling and mechanics and a dedicated electronics and data acquisition system.
EXOGAM will consist of an array of high resolution germanium detectors. These will be arranged to give a high photopeak efficiency of ³ 20% 1.3 MeV gamma-rays. Very high efficiency is required since the beam intensity, at least at the start up of the SPIRAL facility, is expected to be much lower than with stable beams, a factor of 100 or even 1000 lower in many cases of interest. Indeeed the physics interest will naturally be in the study of the most exotic species hence will require the weakest beams.
Segmented CLOVER detectors will be used in the array to give the optimum coverage approaching 4p and optimum performance in a radioactive beam environment. Segmentation is required to provide the optimum performance in terms of efficiency, energy resolution for g ray from recoiling reaction products, and minimisation of multiple-hit events. These Ge detectors will be separated by bismuth germanate BGO suppression shields. They will also have passive heavy metal collimators. Shielding will be required to reduce the background recorded in the active detectors.
An electronics system is needed to process energy and timing information, make trigger decisions and relay data onto an acquisition system. The electronics will be based on the VXI and VME electronics systems in which the UK and France have expertise. Purpose built VXI cards for the new detectors can be made in both the UK and France. A dedicated data acquisition system is needed for EXOGAM which will be based on those developed for Eurogam and Euroball and that currently used at GANIL. It will comprise a DT32 bus output from the VXI electronics, an event builder using VME processors, a VME tape server and several workstations for event monitoring and sorting. The data acquisition system will be capable of recording data up to 2 Mbyte/s.
A system is required to supply the high voltage to the photomultiplier tubes on the BGO detectors and to each germanium detector. This will be based on commercially available hardware and a VME based control system which will be developed by the collaboration. The autofill system to keep the Ge detectors cold can be based on the Eurogam/Euroball systems.
It is envisaged that the array will be used with several ancillary detectors as their use in association with the gamma-ray array will be vital to obtain efficiently the data necessary to achieve the physics goals. A major ancillary detector is the proposed high resolution large acceptance spectrometer VAMOS (the VAriable MOde high acceptance Spectrometer). This spectrometer imposes additional design considerations for EXOGAM since the array may be required to rotate.
Other ancillary detectors include charged particle detector arrays and neutron detectors. A number of such devices are already in existence and will be adapted for use with EXOGAM and several others are being designed. Although the specification, capital and effort required on ancillary detectors is not part of the Exogam project definition, coordination is needed to ensure design compatibility. This work is performed within the EXOGAM ancillary detector working group.
A full description of the detectors and design of the array along with the design considerations in given in section 2. The electronics and data acquisition system is discussed in detail in section 3 and some of the ancillary detectors are discussed in section 4.
2 EXOGAM DETECTOR AND ARRAY DESIGN.
This section summarizes the design considerations for EXOGAM, the types of detectors to be used and the array geometries which are extensively discussed in a specific document: the project definition. The latest advances in germanium detector technology are discussed. The concepts of efficiency, resolution, peak to total, timing response and shielding are introduced and the performance for some configurations of detectors for a selection of typical reactions is presented.
Radioactive beam spectroscopy presents new demands on the design of a g-ray spectrometer. The beam intensity, at least at the start up of the new facility, is expected to be much lower than with stable beams, factor of 10 or even 1000 lower. EXOGAM must therefore be designed to maximise the total photopeak efficiency. This demand of very high efficiency must be achieved for both low (x-ray energies) and high g-ray energy (i.e. 5-6 MeV). Such high energies will be observed, for example, in Coulomb excitation of light nuclei or b-decay studies.
Whatever the type of nuclear reaction, the interest of the physicists is always in the very rare events. In the analysis it must be possible to extract tiny peaks from a possibly huge background (correlated Compton background; background from the radioactivity from the target; scattered beams within the target chamber; etc.) To achieve this it is vital to have a very good peak to background ratio and energy resolution.
The third fundamental criterion specific to the use of a multidetector array with radioactive beams is its modularity. There will be a large variety of nuclear reactions using radioactive beams on which the design of a detection system must be based. The experimental conditions will be very different from one experiment to another in terms of g-ray energy (from x-rays of tens of keV to g-ray energies up to 5-6 MeV), of multiplicity (from one to ~ 15 coincident photons); of recoil velocity (from zero to ~ 10 % of light velocity); and of kinematics of the reaction mechanism (from recoiling fusion products emitted at ~ 0° or scattered particles between 0° and 180°). This variety means that the setup of the array must be adapted for each experiment. This will also be a strong constraint not only on the choice of the individual detector but also on the mechanical structure to maintain the high photopeak efficiency. The radioactive nature of the beam is also a concern and shielding of the detectors becomes an important design criterion.
It is also clear that in addition to the detection of gamma radiation it will be vital to have ancillary detectors available to detect both light and heavy charged particles and neutrons.
2.2 Gamma-ray detector developments.
In order to meet all the design criteria the EXOGAM spectrometer will consist of an array of high resolution germanium detectors each surrounded by an escape suppression shield usually made from bismuth germanate (BGO). The development of such escape suppressed spectrometers has revolutionised the field of gamma-ray spectroscopy over the past 15 years [3.1-3.3]. These arrays are designed to measure efficiently the energies of and the correlations between cascades of gamma-rays depopulating nuclear states. These arrays involved significant technological advances in detectors, electronics and data acquisition.
The main design criterion for EXOGAM is that it must have a high detection efficiency. However, in maximising efficiency the spectrum quality must be maintained. The spectrum quality is determined by the peak to total ratio and energy and time resolution. The total efficiency measures the ability of the array to collect statistics. The spectrum quality measures the effectiveness of the array in isolating a single sequence or sequences of gamma-rays from a complex spectrum. The arrays performance can also be improved using ancillary detectors in coincidence with the array.
Recent detector developments have been made that enable high efficiency detectors to be made which maintain excellent spectrum, quality. These detectors are segmented, composite detectors. A composite detector is made up of several Ge crystals packed closely together in the same cryostat. In this way a large detector can be created with high photopeak efficiency and high resolving power, since the (small) individual crystals minimise the effect of the Doppler broadening. An array of such detectors has a high granularity and high efficiency. The first composite detector to be used in a large array is the CLOVER detector [3.4]. The CLOVER detector consists of four co-axial n-type Ge crystals arranged in the configuration of a four leaf CLOVER and housed in the same cryostat. The individual crystals for the EXOGAM Clover will be 60 mm diameter and 90 mm long before shaping. Adding the signals corresponding to scattered events between adjacent crystals enhance significantly the efficiency. The gain in efficiency obtained in this way compared with the sum of the individual crystals is called the addback factor. Using these detectors at ~ 90° to the beam direction significantly reduces Doppler broadening. The granularity, and hence the Doppler broadening can be improved still further by electronically splitting the crystal. The technique of electronically segmenting the outer surface of a n-type crystal has recently been developed and it is now possible to combine composite and segmented detectors.
2.3 Segmented CLOVER Ge detector
One composite and segmented detector is the segmented CLOVER detector. The segmented CLOVER detector has each of the individual crystals electronically segmented into four regions. A schematic diagram of the crystals in a segmented CLOVER detector is shown in figure . The EXOGAM array will consist of segmented CLOVER detectors.
Figure 1: The segmented CLOVER germanium detector crystals.
This segmentation is particularly useful when the emitting nuclei recoil has a large velocity since it allows a better determination of the interaction point of the g ray in the detector. The detectors can be easily arranged in different configurations as will be shown later. This high degree of versatility is a very important design criterion as in the future the full gamut of nuclear reactions with stable and radioactive beams will be used for nuclear spectroscopy.
Four good energy resolution signals ( < 2.1 keV for 1.3 MeV g-rays) are output from the four centre contacts. Lower resolution signals are output from 16 outer contacts and these enable the position of the event to be determined. The in-beam resolution is significantly improved because of the reduced Doppler broadening effects. This segmentation is especially suited for very close geometries or nuclear reactions, for which the gamma-ray emitting nucleus has a high recoil velocity and/or is scattered with a broad range of scattering angles.
Characteristics of the EXOGAM segmented CLOVER
GEANT simulation calculations have been carried out to optimise the performance of CLOVER detectors for EXOGAM. The EXOGAM CLOVER will be based on the use of large Ge crystals, 60 mm in diameter and 90 mm long, before shaping. The increase in the intrinsic photopeak efficiency and diameter allows for a compact design maximising the solid angle coverage of Ge. The length ensures a large efficiency up to the highest g-ray energies. The calculations have shown that a tapering over a length of ~ 3 cm of the crystals optimises both the efficiency and the peak to total.
The calculated effective photopeak efficiency and peak to total as a function of g-ray energy and multiplicity are shown in figures and , respectively for the EXOGAM segmented CLOVER.
Figure 2: Calculated effective photopeak efficiency and unsuppressed peak to total as a function of gamma-ray energy at a distance of 11 cm for the EXOGAM CLOVER.
Figure 3: Calculated effective photopeak efficiency at 1.3 MeV and peak to total as a function of gamma-ray multiplicity at a distance of 11 cm for the EXOGAM CLOVER.
Pile up effects.
When more than one g-ray is emitted simultaneously from a source, it is possible to detect more than one of them simultaneously in a given CLOVER. This probability increases rapidly with the multiplicity and the absolute (and not only photopeak) efficiency. This pile up effect induces a loss in photopeak efficiency and degrades the peak to total ratio. Calculations have shown that at ~ 11 cm the effect of pile up is too severe when the multiplicity is larger than Mg ~ 15. This is demonstrated in figures and which show the total photopeak efficiency and the addback factor, respectively, at 1.3 MeV as a function of g-ray multiplicity for the 16 segmented CLOVER array in configurations A (Ge crystal to target distance d = 11.4 cm) and B (d = 14.7 cm), see section 3.6.
Figure 4: Total photopeak efficiency for the 16 segmented CLOVER array in configurations A and B at 1.3 MeV as a function of g-ray multiplicity.
Figure 5: Addback factor for the 16 segmented CLOVER array in configurations A and B at 1.3 MeV as a function of g-ray multiplicity.
The great advantage of the segmentation is that the opening angle subtended by individual segments is reduced by a factor of two compared with the one of the crystals. The opening angle is a major contribution to the Doppler broadening of the photopeaks for high v/c which means that the subsequent reduction in the photopeak width can be up to a factor of two. This improvement is fundamental when the background is large, and in particular when using radioactive beams. A comparison of the calculated resolution for a single crystal detector, and a segmented and non segmented CLOVER is shown in figure . The calculation is for detectors at 90°, 6 cm from a source moving with a recoil velocity of v/c = 7.5%.
Figure 6: The calculated energy resolution D Eg as a function of gamma-ray energy for a single crystal detector, and a segmented and non segmented CLOVER. The detectors are at 90°, 6 cm from a source moving with a velocity of v/c = 7.5%.
Another improvement on the energy resolution can be obtained by analysing the shape of the pulse delivered by the preamplifier. Indeed, this shape is characteristic of the distance between the interaction point and the anode of the Ge diode. A substantial gain will however be obtained only if the hit localisation is precise enough and tests will be performed in the near future to estimate this gain. Information on the pulse shape for radial position determination will be included in the VXI cards on the signal from the central contact.
In addition, information on the pulse shape will improve the timing performance of the detector. This is extremely important in radioactive beam experiments since the uncorrelated background radiation can be greatly reduced by precisely relating the events in the detector to the beam. This was demonstrated in the g-ray spectroscopy experiment using the radioactive 19Ne beam at Louvain-la-Neuve [3.5]. The use of the time relationship between the pulsed beam and germanium detectors was vital for the success of the experiment.
Calculations of the Rutherford scattering cross-section have been performed for several 'typical' reactions using radioactive beams, see section 3.7. Figure shows the differential Rutherford scattering cross section as a function of scattering angle for a series of reactions. Assuming this radiation is deposited in the target chamber wall and is in the sight of the detectors, then the resulting background rate can be calculated. For an average background multiplicity of 2 a rate of > 2 kHz will be recorded in the Ge detectors for all these reactions if the detectors are in the region of £ ± 30° relative to the initial beam direction.
Figure 7: The differential Rutherford scattering cross section as a function of scattering angle for a series of typical reactions, see section 3.7.
2.4 Suppression shield for a segmented CLOVER Ge detector
Each segmented CLOVER Ge detector is surrounded by an escape suppression shield. The shield designed for the EXOGAM CLOVER is based on a new concept in which the shield comprises several distinct elements, a backcatcher, a rear side element and a side shield, see figure . Designing suppression shields in this way, from individual elements, creates greater flexibility for different configurations.
Figure 8: The different elements of the BGO suppression shield for the segmented CLOVER Ge detectors (not to scale).
The shields will be operated in two configurations. The first is with the back catcher and rear-side element, configuration A, and the second with the additional side elements, configuration B. The only free space is left for the cold finger from the liquid nitrogen dewar of the Ge detector.
The use of a back catcher prevents forward scattered events from escaping through the large angular section located behind the CLOVER. The calculations have shown that the use of the back catcher will improve the peak to total by ~ 10%. The energy of forward-scattered g-rays are the highest since they deposit very little energy in the Ge crystal. This implies that a larger thickness of BGO is required. Simulation calculations have shown that at 1.33 MeV about 95 % of the maximum possible peak to total value can be obtained when this thickness is ~ 2.5 cm at the rear and the width of the rear side element ~ 2 cm. The thickness of the side element must be at least 1 cm. To maximise the solid angle coverage with Ge material (i.e. the photopeak efficiency), the side elements of the shield have be in BGO. However, behind the Ge detector there is sufficient space for a scintillator with lower stopping power to be used, such as CsI(Tl). The backcatcher can therefore be CsI(Tl), which will give the same overall suppression performance at lower cost.
Figure shows the calculated gain in peak to total as a function of g-ray energy compared with an unsuppressed CLOVER. The gain for the full suppression shield and for just the rear side and back catcher elements of the shield is shown.
Figure 9: Calculated gain in peak to total ratio as a function of g-ray energy compared with an unsuppressed CLOVER. The blue line is for a suppression shield with the rear side elements and back catcher, configuration A. The red line is for the full suppression shield including the side crystals, configuration B.
Each part of the suppression shield will comprise several (4-8) bismuth germanate BGO / CsI(Tl) crystals. The outputs from the photomultiplier tubes will be coupled together such that 4 signals are presented to the electronics. The signals will be coupled together in order that the appropriate suppression elements can be used to suppress an individual CLOVER Ge crystal.
A heavy metal collimator is needed for each escape suppressed spectrometer to prevent direct radiation from the target reaching the suppression shield (configurations B) and to prevent cross scattering in the array (configurations A.)
The EXOGAM segmented CLOVERs can be arranged in different geometries. In all the geometries the suppression elements can be used configurations A and B.
Configuration A is the close packed geometry where the Ge detectors can essentially touch at the front. In the design of configuration A, 0.5 mm clearance is allowed between the housings of the Ge detectors. Configuration B is the pulled back geometry in which the detectors are further from the target to allow for the inclusion of the additional side suppression elements. In this configuration 26 mm is allowed between the Ge detectors for the two BGO detectors. This clearance includes the two 10 mm thick BGO crystals, the crystal cans and clearances.
An array geometry for the CLOVERS to be as close as possible to the target is with the detectors on the faces of a cube. In this geometry the Ge crystal are at 50.3 mm from the target in the close packed configuration A and 68.3 mm in a pulled back configuration B.
An array of 16 Clover detectors can be arranged with 4 detectors at 135°, 8 detectors at 90° and 4 detectors at 45° to the beam direction. A cross section through this geometry in configuration A is shown in figure and an isometric projection in figure . In configuration A the signals from adjacent Ge crystals can be summed to increase the efficiency. The calculated increase in efficiency is 6%. This geometry leaves space for beam in and out. The front face of the Ge crystals is 114.1 mm from the target in configuration A and 147.4 mm in configuration B.
Figure 10: A cross section through the 16 segmented CLOVER EXOGAM array.
The configurations and distances from the target for arrays of segmented CLOVERs are summarised in table 1.
|Geometry||Suppression shield |
|Distance to the |
|16 detector geometry||A||114.1|
|16 detector geometry||B||147.4|
In the 16 detector geometry the CLOVERs in the forward angle position (22.5° to 67.5°) will have a high background rate from Rutherford scattered beam and may have to be removed or collimated for certain experiments.
An alternative to this solution is to replace the 4 forward-angle CLOVERs with a ring of 8 segmented single crystal detectors. The 8 single crystal segmented detectors each have a crystal diameter 70 mm with a 30 mm taper at the front and are segmented both on the outer and inner contacts. The segmentation of both contact of an n-type Ge detectors is a recent technical development. The length of the crystal be 90 mm. The crystals are housed in a cryostat of diameter 88 mm. It is possible to locate 8 such detectors at 50.4° with the crystal to target distance 98.2 mm. This is the close packed configuration A for these detectors. These detectors have to be pulled back to 43 mm for a full suppression shield, configuration B for these detectors. These detectors are outside the ±30° region and therefore should not have too high a background rate from Rutherford scattered beam.
[3.1] J.F.Sharpey-Schafer and J.Simpson, Progress in Particle and Nuclear Physics. 21 (1988) 293-400.
[3.2] P.J.Nolan et al.,
[3.3] C.W.Beausang and J.Simpson, J.Phys. G. 22 (1996) 527-558.
[3.4] F.A.Beck et al., Proc. Workshop on Large g Ray Detector Arrays, (Chalk River), Canada, AECL-10613, p 359.
[3.5] W.Catford et al., Nucl. Instrum. Meth. A 371 (1996) 44.