The PAMELA calorimeter is a sampling calorimeter made of silicon sensor planes interleaved with plates of tungsten absorber. The calorimeter performs a precise measurement of the total energy deposited, reconstructs the spatial development of the shower both in the longitudinal and in the transverse directions and precisely measures the energy distribution along the shower itself.
The main physics tasks of the imaging calorimeter are the following:
- extraction of the antiproton signal from the large electron background, with an efficiency of about 90 % and a rejection power of 10-3 - 10-4 ;
- identification of positrons in a vast background of protons with an efficiency of about 90 % and a rejection power better than 10-4.
General characteristics
To accomplish the above mentioned tasks, the PAMELA calorimeter has to have a high granularity, both in the longitudinal (Z) and in the transversal (X and Y) directions. In the Z direction, the granularity is determined by the thickness of the layers of the absorbing material (tungsten); each tungsten layer has a thickness of 0.26 cm, which corresponds to 0.7 X0 (radiation lengths). The total depth is 15.3 X0 (i.e. 0.9 interaction lengths), since there are 22 layers of tungsten.
Each tungsten layer is sandwiched between two layers of silicon detectors, i.e. the stratification of a single plane is Si-X/W/Si-Y. For each view (X or Y) there are 9 silicon detectors, arranged in a square matrix of 3x3 detectors. Since each silicon detector has a surface of 8x8 cm2, the total sensitive area is 24x24 cm2. The total sensitive volume is 24x24x18 cm3. Each detector has 32 strips and each strip is connected to the corresponding ones belonging to the two detectors of the same row (or column), so that the number of electronics channel per plane is 32 x 3 x 2 = 192 and the total number of channels is 192 x 22 = 4224.
Mechanics
The mechanical structure is based on a modular concept. The basic unit is called a "detection plane", and it consists of an absorber plate, two PCBs (X and Y, supporting the silicon detectors as well as the fron-end and part of the read-out electronics) and the two matrices of silicon sensors (the first and last detection planes have only one layer of silicon sensors). Two detection planes form a "detection module". In a module, the two detection planes are kept together by a frame to which they are bolted at the edge of the absorber plate (Figure 1). The 12 modules are independent and fully extractable; they are inserted like "drawers" in the main mechanical structrure and then locked by a cover (Figure 2).
The total calorimeter mass (including electronics and cables) is 110 kg. The whole calorimeter structure has been modelled and numerically analysed by means of the finite element method.
Silicon detectors and interconnection technique
The interconnection technique to realize the detection planes have been defined through numerous tests, in collaboration with the firm Mipot (Cormons, Italy), using both "mechanical" and "real" silicon detectors. As a first step, the printed circuit boards are fixed to the corresponding tungsten plates. The detectors are glued, in rows of 3, on a specially designed 75 mm thick silicon detectors). thick kapton layer with a siliconic glue. Then, the wire bonding of the corresponding strips on each detector is performed. Afterwards, 3 such rows are glued on the supporting printed circuit board, again by using a special siliconic glue, to form the 3x3 silicon detector matrix of one view (either X or Y).
The silicon detectors for the PAMELA calorimeter are very large area devices (8 x 8 cm2 each), 380 mm thick and segmented into 32 strips with a pitch of 2.4 mm. They feature an innovative bias technique, adopted in order to bring the bias voltage with a wire bonding directly on the junction side of the devices. To accomplish that, the bulk contact is realised via a forward biased p+ implant running at the edge of the device. In this way, we can avoid the use of a conductive epoxy glue for gluing the detectors on the PCB, therefore preserving the devices from the large mechanical stress that these type of glues can induce during polymerization.
Several pre-series detectors, from two manufacturers, have been completely characterised in the laboratory. The test results are very good: the average value of strip leakage current is about 400 pA, which corresponds to an average current per unit area of 0.17 nA/cm2. Figure 3 shows the result (for one strip) of an I-V measurement performed over 72 hours on a pre-series detector. As one can see, the stability of the current, which is a crucial item in space applications, is excellent.
Front-end and read-out electronics
The front-end electronics is based on a VLSI ASIC: the CR-1 chip. The use of an ASIC allows to gain considerably in weight reduction and compactness with respect to the discrete preamplifiers previously used in balloon flights. The main design characteristics of this chip are the very large dynamic range (1400 MIPs), the ability to cope with a very large (about 180 pF) detector capacitance, the good noise performance (2700 e- rms + 5 e-/pF) and the low power consumption (< 100 mW/chip). Each circuit has 16 channels and each channel comprises a charge sensitive preamplifier, a shaping amplifier, a track-and-hold circuit and an output multiplexer. A self-trigger system and an input calibration circuit are also integrated on chip. Figure 4 displays a scope picture of the output of a CR-1 circuit (trace 2) when stimulated by an input calibration signal (trace 1) injecting a charge corresponding to 300 MIPs (Minimum Ionizing Particles, 1 MIP 5.1 fC for 380 mm thick silicon detectors).
The design of the read-out electronics is divided into two main parts: ADC electronics (on the detector boards) and Data Processing electronics (on external read-out boards).
On each detector board, the 6 CR1 outputs are connected to a 16-bit ADC (with serial digital output) through an analogue multiplexer and an operational amplifier. On the read-out boards the collection and analysis of the events, prior to their transmission to the main CPU, is performed. The whole calorimeter is divided, from the point of view of the read-out, into four independent sections. An FPGA parallelize the data of the ADCs, controls the generation of the multiplexer address and performs the self trigger coincidence. Four DSPs (Digital Signal Processors, one for each section of the calorimeter), read and process the data and control the acquisition procedure.
Self-trigger calorimeter
The calorimeter will be also used in a self trigger mode to increase the geometric acceptance and to identify high energy electrons and photons. For more details see the postscript file (6.6 Mbyte).
Contacts
For further information, contact
- Valter Bonvicini
- Mirko Boezio
- Tel: +39 040 375 6224; Fax: +39 040 375 6258