Germanium crystals have long been used to study photons with energies from 50 keV to 10 MeV. Their excellent energy resolution (approaching 0.1%) has created numerous applications in nuclear and particle physics, especially in studies of nuclear structure. Their major limitations are their poor position resolution and inability to reconstruct multiple interactions. Now, germanium crystals are being made to do “double duty”, measuring the interaction points as well as the deposited energy, which allows for full 3D reconstruction of the energy deposition.
Photons with energies of less than a few million electronvolts interact primarily by Compton scattering. They usually interact several times before stopping, and many photons escape from conventional detector arrays without depositing their full energy. These partially reconstructed events constitute a substantial background to measurements. To reduce this background, existing germanium detectors are usually surrounded by thick anticoincidence (veto) counters. This veto greatly reduces the efficiency of large detector arrays.
The new breakthrough is to make germanium crystals work like miniature time-projection chambers, with the charge deposition measured at each point in the crystal. A central cathode embedded in the crystal generates a radial electric field. Electrons liberated by photon interactions in the crystals drift to segmented anodes that cover the crystal surface. Charge sharing between adjacent electrodes allows position resolutions of 1-2 mm, far better than the current one-crystal (5-10 cm) resolution. The electron drift time is also measured, which gives the depth of the interaction in the crystal and provides 3D space points. With good segmentation, complex interactions can be reconstructed, which greatly increases the photon detection efficiency while maintaining optimum resolution.
In many experiments to study very unstable nuclei, excited nuclei are produced at high velocities. To obtain gamma-ray spectra from these nuclei it is necessary to correct the photon energies for the nuclear Doppler shift; the accuracy of this correction depends on the precision of the photon position measurement. Another important application is precision nuclear spectroscopy, where the increased efficiency is needed to study multistep decays. For example, highly spinning nuclei may emit 20 or more photons as they de-excite. The ability to detect many photons in a single event greatly increases the experimental sensitivity to these reactions; high efficiency is critical for obtaining the required high-coincidence spectra.
Two large collaborations are developing gamma-ray tracking arrays using segmented crystals with appropriate read-out. In the US, the Lawrence Berkeley National Laboratory-led GRETA/GRETINA collaboration is building a segmented triple-crystal prototype module. Each crystal is covered with 36 electrodes (see figure). The read-out electrodes are segmented longitudinally and transversely. Each channel is instrumented with a low-noise pre-amplifier and a fast (100 Megasamples per second), accurate (14 bit) analogue-to-digital converter. The energy resolution is 1.9 keV for 1.33 MeV gamma rays, which is comparable to the best unsegmented detectors. GRETINA will be composed of 10 triple-crystal modules covering about 25% of 4π. It will travel from accelerator to accelerator, following the best physics. The follow-on to GRETINA, the 120-crystal GRETA detector, will have full 4π coverage.
The proposed 180-crystal (6500 channels) European AGATA array, also for nuclear spectroscopy, uses a similar technology to GRETINA. These arrays will have figures of merit that are several orders of magnitude better than existing large arrays, such as Gammasphere at Argonne and Eurogam at IReS in Strasbourg.
A few smaller arrays are already operational. At the Michigan State University cyclotron in the US, the SeGA array comprises 18 crystals, each with 32-segment read-out. These crystals are slightly smaller, with a 5 keV energy resolution. The EXOGAM array at GANIL in Caen, France, has 64 crystals, each with four segments, to measure the depth of interaction. Similarly, Miniball at CERN has 40 crystals with six segments. The proposed Canadian TIGRESS array at TRIUMF will comprise 64 eight-segment crystals.
Even for simple events, the improved position resolution is an important development. The resolution could lead to better images from positron emission tomography cameras, where two reconstructed 511 keV photons are used to localize positron annihilation in patients for various medical and biological applications.
The technique may also be used to reduce backgrounds in double beta decay and dark matter searches. The US Majorana collaboration proposes to build a 200-crystal germanium detector containing 500 kg of 86% enriched 76Ge to study these topics. Simulations indicate that the position resolution obtainable with segmentation can lead to a factor of 5 to 8 rejection in backgrounds.
Further reading
For further information on some of these projects see: http://greta.lbl.gov/
