The gamma camera uses a double-cone collimator to form a visible-light or gamma image in the same way as a pinhole camera. Unlike the initial design, the image is not formed on a photographic emulsion, but on a scintillator: a CsI(Tl) disk 4 cm in diameter and generally between 2 and 4 mm thick. Gamma photons interact in the scintillator to generate luminous photons with a wavelength of 550 nm, forming an image. As in medical imaging systems, the scintillator output signal is amplified and collected on a CCD array. We use a second-generation image intensifier tube with a double microchannel plate, and a large diameter (40 mm) entrance window. The CCD array measures less than a centimeter on a side, and is therefore coupled to the tube screen via a tapered fiber-optic demagnifier. The unit is compact (12 cm in diameter and 30 cm long) and is protected against spurious radiation by Denal tungsten alloy shielding 2 cm thick, ensuring 85% attenuation at 1.25 MeV and 95% at 660 KeV. A color video camera is secured to the shielding.

The gamma camera and color camera signals are transmitted to video acquisition cards in a PC-compatible microcomputer running the imaging software.

The sensitivity of the Aladin 2 device allows "real-time" detection of hot spots (i.e. within 5 seconds). In this mode, the image acquisition is performed entirely on the CCD array, which accumulates the incident signals for a predetermined time (t = k ´  40 ms); this is equivalent to setting the exposure time of a conventional camera. The resulting accumulated signal is transmitted to the PC video acquisition card; the screen displays an 8-bit image updated every k ´  40 ms.

The facilities examined are often large and complex, and the hot spot locations are not known beforehand. The "detection" mode allows the camera to sweep the field of view while the PC monitor displays the real-time color video image and the periodically updated gamma image. A hot spot results in signal saturation, producing a "white spot" that is clearly distinguishable from the noise; it does not provide precise information on the source location, but is conclusive evidence of the existence of an irradiating zone.

The least favorable case would be a facility in which the hot spot dose rate at the camera distance is very low, on the order of 10 m Gy.h^-1 (1 mrad· h^-1). Under these conditions the CCD integration time is 5 seconds, requiring slow video sweeping. In most case, however, the irradiation levels generated by hot spots range from 1 mGy.h^-1 (100 mrad· h^-1) to 1 Gy.h^-1 (100 rad· h^-1) at the camera position: detection is possible in only a few tenths of a second or even in real time (40 ms). Figure 1 shows the image corresponding to in real time of a hot spot irradiating a around 0,4 Gy.h^-1 at the camera position.

Once a hot spot has been detected, an 8-bit dynamic range is rarely sufficient to localize the source without accumulating a large number of incident images. The video acquisition card is used to integrate successive CCD signals (generally in real time) into a 16-bit image; white spots appear on the display with increasing intensity during the integration period. The acquisition time ranges from 5 seconds to 10 minutes, depending on the irradiation level and on the intensifier tube gain. Figure 2 shows a typical integrated gamma image of a highly irradiating source with a camera-position dose rate of 4 Gy.h^-1.

In its present configuration, the device is capable of detecting point sources irradiating a 10 m Gy.h^-1 (1 mrad.h^-1) dose rate, in 5 seconds (8-bit image depth). Localization of a source generating 0.5 m Gy· h^-1 (0.05 mrad.h^-1) requires a 10-minutes exposure (Signal to Noise ratio » 2). The photon flux is very weak for low-level irradiation sources : for example, a dose rate of 10 m Gy· h^-1 (1 mrad· h^-1) generated by ⁶⁰Co photons (1.25 MeV), corresponds to a photon flux of 520 ph· cm^-2s^-1. The limiting factors are the aperture diameter of the collimator (1.4 mm is used to obtain an angular resolution of a few degrees), and the poor interaction probability (4.4 ´  10^-2 interaction per photon for 1.25 MeV photons trough a 2 mm CsI(Tl) scintillation crystal). The detection sensitivity can be improved ,at the expense of the resolution, through the use of collimators with a larger aperture. Another option would be to use a thicker scintillator crystal to increase the interaction probability.

The angular resolution varies according to the photon energy flux, the aperture diameter, the scintillator thickness and the collimator solid angle. The present configuration (1.4 mm aperture, 2B4 mm thick scintillator, 50E collimator angle) is capable of discriminating ⁶⁰Co sources with an angular separation of less than 5E (the separability criterion assumes 30% relative contrast between the two peaks). In the most favorable case (cesium) the resolution is about 2.5E. An effective way of enhancing the resolution is to diminish the collimation angle: a 30E collimator provides an angular resolution of 1.3E at the same energy, but with a corresponding reduction in the field of view.

Detection of hot spots in a few seconds can be useful not only in decommissioning a nuclear facility, but also during routine operation or maintenance, for instance to follow a moving source. Localizing irradiating source in a few minutes enables multiple views from different angles to be observed, for accurate localization, especially in complex environments.

After precise calibration, it is possible to estimate the dose rate generated by a punctual source at camera position. As a matter of fact, the pixel size reached after a given integration time depends on the dose rate received (for given amplification gain and photon energy). The knowledge of dose rate 1 meter apart from the source requires the use of a telemeter. Once the source has been localized, more precise data on the source (accurate dose rate, gamma spectrum) may also be obtained by the use of a strong collimated separate probe.

Strongly contaminated a sources can be localized in a relatively short time: only one minute was necessary to obtain an "alpha" image of an americium-241 source with a contamination level of 100 MBq (3.3 MBq.cm^-2) at a distance of over four meters (S/N ratio » 4). This result was obtained in darkness with a relatively narrow field of view (6° solid half-angle). At the detection limit, a plutonium point source generating 1350° Bq over 1° cm² at a distance of 4° m can be hardly detected, through a 5 mm plexiglas wall, in 30 minutes. The same source is easily detected at a shorter distance, 1.2 m, without plexiglas screen, in 15 minutes (S/N ratio » 9).

Tests also showed that alpha sources can be observed through Plexiglas, triplex or polycarbonate screens of limited thickness. The attenuation coefficients of these three materials were found to range from 0.85 to 0.97 per millimeter.

The results of a test under realistic conditions (a telemanipulator tong observed from a distance of 1.25 m through a 5 mm Plexiglas glove box) are shown in Figure 4. The major contaminants were Curium isotopes at levels of around 1 MBq· cm^-2, allowing short acquisition times on the order of five minutes. The most severely contaminated areas are clearly visible: mainly the inside of the jaws and a few edges or asperities around the body.

It is interesting to observe the area between the jaws, where scattered and attenuated points are visible. This may be attributable to the medium-level energy (about 6 MeV) of the a particle emitted by the curium isotopes. Alpha particles follow straight-line paths in air, with a mean stopping power of 1.3 MeV$cm^-1. Thus the range of alpha radiation with an energy of 6 MeV reaches 4.3 cm, compared with the 8 cm gap between the jaws.