Beam diagnostics for low-intensity radioactive beams

In order to have a suitable check of the beam properties (profile, intensity, ion composition, etc.) for the beam tuning requirements, EXCYT needs an efficient beam diagnostics. The foreseen low beam intensity (tipically below 10⁸ particles/s) does not allow the use of classical devices based on the measurement of the carried charge (wire devices), since the produced currents (< 100 pA) would be too low to be efficiently discriminated from the background noise (~10 pA). Therefore the adopted solution consists to adopt particle detectors which, being sensitive to the energy released by the particles, tipically have a better sensitivity. Our research and development activities have led to the construction of a series of devices based on particle detectors, such as semiconductors and scintillators, that have shown how the radioactive beam should be handled and identified both at low (50 keV to 300 keV) and high (up to 8 MeV/A) energy.

Preacceleration beam imaging and identification. The LEBI device

The beam diagnostics in the preacceleration stage (low energy) along the beam pipe, between the target ion-source complex and the Tandem accelerator [1] is a crucial point, since a quick beam tuning needs an efficient real-time check of the beam properties. The diagnostic devices should be able to locate the beam position, to measure its transversal size and to identify its nuclear composition. Unfortunately that is a difficult task, because of the low energy and intensity of the radioactive beam. However the space separation between the different isobaric components, which depends on their mass excess, can range from a few millimeters, in the most favourable cases (light species), to a few hundred microns in the worst ones (heavy species).

High luminous efficiency CsI(Tl) scintillating plates have been successfully used for beam imaging and profiling, for applications inherent in stable and radioactive beams. Imaging of stable beams is obtained when ions directly hit the plate surface. The released energy is converted into scintillation light, so that a light spot representing the transversal profile of the beam is produced. A CCD camera watches the light spot, which is directly observable on a monitor and acquired by a PC. For low energy radioactive beams we cannot exploit the kinetic energy of ions, because also of the expected low beam current, and besides the radioactive contamination of the plate has to be prevented. The solution is to exploit the radioactive decays produced by the radioisotopes of the beam (mainly b and g rays), implanted in a thin inert tape placed in front of the scintillating plate. Starting from this working principle, the device named LEBI (Low Energy Beam Imager / Identifier) has been designed. It permits to attain the beam imaging and identification by exploiting the radiation emitted by the radioactive ions, fig. 1, and basically consists of a CsI(Tl) plate, 1 or 2 mm thick, and of a 6 mm thick mylar tape arranged in front of the plate, at a distance of 0.5 mm [2,3].

Fig. 1. Sketch of the LEBI device for low energy beam imaging and identification.

When the film and the scintillator are placed along the beam line in order to intercept the beam, the ions get implanted onto a small film area (the transversal area of the beam) because of them low energy (up to 300 keV). This area thus becomes a radioactive source, in which the activity due to each kind of isotope implanted is described from the following well known exponential law, (t = 0 sec when the beam starts):

,

where r is the beam rate of each radioisotope and l is the related decay constant.

The radiation is emitted isotropically, and since a large fraction of the solid angle covered by the plate is close to the emission point, the sum of the contributions due to each beam particle gives rise to a light spot, whose diameter is slightly larger than the transversal size of the beam. The brightness of the light spot is related to the activity of the implanted nuclei and to the type and energy of the radiation produced by the decay. If r is larger than 10³ pps and the mean lifetime t=1/l is not longer than a few seconds, the beam will be reasonably observed shortly after the beginning of the implantation. Conversely, if the rate is lower and/or t is too long (>1000 sec), the observation could be troublesome even by using a high sensitivity CCD camera. The spatial resolution of LEBI is rather modest, mainly because the radiation is emitted isotropically. So, if we use a hypothetical point like source placed in front of the plate, the radiation crosses the plate in all directions (the plate covers a solid angle of about 2p sr), thus producing a light spot with a halo around it. The FWHM of the spot profile represents the spatial resolution of the system, which is of the order of the plate thickness, in our case between 1 mm and 2 mm. Since no radioactive beam is still available from EXCYT, an experimental test has been performed by using a 1 mm collimated ⁹⁰Sr beta source (the decay chain produces two beta particles, with endpoint energy of 546.2 keV and 2280.1 keV respectively), with outgoing beta rays intensity of 10³ pps. It was placed in front of the 2 mm thick CsI plate, at a distance of 1 mm from it. The well visible light spot obtained with this plate is shown in fig. 2.

Fig. 2. Light spot of the 1mm ⁹⁰Sr beta source. The FWHM is 1.7 mm and taking into account the collimator diameter, 1 mm, the estimated spatial resolution is 1.5 mm.

The transversal profile has been fitted with a gaussian curve with FWHM » 1.7 mm, so that by using the rule of the sum of the squares, we calculated a spatial resolution of Dx » 1.5 mm.

In order to get as much information as possible to identify the beam, a small photomultiplier (Hamamatsu R7400) used in pulse counting mode is optically coupled to a side of the plate, by means of a light guide. Its main application concerns the identification of implanted nuclear species, by measuring the particle count rate at fixed time intervals, in order to estimate the decay constant l. For decays in which the daughter nuclei emit gamma ray, a couple of high purity germanium detectors installed close to the plate, allow a most suitable identification of the isotopes. Since the gamma ray spectrum is typical of each nucleus, the recognition of well defined peaks by means of gamma spectroscopy, allows the identification of the different nuclear species present in the beam. The germanium detectors for gamma identification are positioned very close to the mylar tape, at a relative angle of 90°. They should collect events with at least two gamma rays emitted in coincidence, so that the background can be strongly reduced, highlighting the gamma cascades bound to the selected gammas. In such a way it is possible to perform a strong selection of the nuclear species, provided that it has at least a couple of gamma rays in cascade. We have tested this technique with two germanium detectors and a ⁶⁰Co source, showing that it is reliable. fig. 3.

In order to study how LEBI should display the beams transported along the beam pipe of EXCYT, we have developed a Monte Carlo simulation code, based on the energy loss of beta rays inside the crystal. What comes out is that the only contribution relevant for the detection is the beta decay. Therefore we developed a set of ad-hoc tools in order to evaluate the specific energy loss and range of beta electrons/ positrons in matter, together with the BIS++ code that makes a Monte Carlo simulation of the scintillation produced by beta rays. It is capable of simulating the shape of the light spot produced by the radiation crossing the plate. As an example where a realistic beam is simulated, we assumed to produce a ¹⁸F beam that contains ¹⁸N as a contaminant. Using beam transport calculations, we derived the transverse distribution of the two ion species after the mass separator; the foreseen separation between the centroids being 4.8 mm. Then we used these results as input to the LEBI simulation code, whose output is reported in fig. 4. The spatial separation between the main beam and the contaminant is evident. The predominance of the contribution due to ¹⁸N ions depends on the value of its decay constant (l_18N = 1.11 sec^-1), which is much larger than ¹⁸F (l_18F = 1.05×10^-4 sec^-1).

Fig. 3. Gamma spectrum of the ⁶⁰Co source; free (upper) and conditioned to one peak (lower).

Fig. 4. Simulated response of LEBI for a ¹⁸F beam and its contaminant ¹⁸N.

The LEBI device we have built is made of a spherical vacuum chamber containing the plate-tape set-up, fig. 5. The thin mylar tape is rolled up in two spools and can be slid on, whenever it becomes contaminated, by means of a motor. An external high sensitivity CCD camera (sensitivity of 3×10^-4 lux) watches the plate and is connected to a frame grabber for the acquisition by a pc. A pneumatic cylinder allows to insert and remove the plate-tape set-up from the beam line via remote control. The germanium detectors are arranged by using two cups assembled in the vacuum chamber. A set of software tools to be developed, will offer enough flexibility for managing the different peculiarities of each produced beam. At the moment we have programmed to install five LEBI devices in the first stage of the mass separator and four in the second stage.

Figure 5. Detail of the LEBI device. The CsI(Tl) scintillating plate and the mylar tape are evident.

Post-acceleration

Post-acceleration beam profiling

After the acceleration by means of the Tandem (15 MV), the beam energy is often high enough to allow the direct detection of ions by using particle detectors.

The Glass FIbre Based Beam Sensor (GFIBBS) represents our general solution for beam profiling of EXCYT, since we proved it is reliable, cheap and simple [4]. It is based on a pair of Terbium doped glass fibres scanning the beam. The two fibres are mutually perpendicular and are readout by means of a single compact PMT; it allows to reconstruct the X and Y beam profiles in a single scan with high efficiency, figs 6, 7.

In order to develop new beam diagnostics tools, which should be able to cover the wide intensity range of the beams, different techniques have been considered based on gas detectors, secondary emission and scintillators, as described in the following.

Fig. 6. Sketch and picture of the GFIBBS device.

Fig. 7. Extracted profile with GFIBBS of a Nickel beam.

The moving slit sensor, named SBBS (Scintillator Based Beam Sensor) and shown in fig.8, consists of an inorganic scintillator crystal (CsI) placed behind a thick graphite screen with a 0.5 mm slit. The scintillator is optically coupled with a compact photomultiplier, and the whole structure can be moved to scan the beam. Quite good results have been obtained in terms of sensitivity, even though the device completely stops the beam while in operation. The same device has also been successfully used to count the single particles of the beam in case of very low intensity, allowing the self calibration (light versus counts). We also proved that this device can operate at very low energy, by easily sensing a 1 pA beam of ¹²C at 50 keV [5].

Fig. 8. The SBBS device.

Gas detectors can also be used for the beam profile reconstruction, where the signal is produced by ionization due to the energy lost by the ions in a chamber filled with a suitable gas, or by interaction with the residual gas present along the beam pipe. In the last case, the very few ionizing collision events need some sort of physical amplification; therefore a microchannel plate (MCP) is generally used, onto which the incoming electrons or ions produced by ionization are driven by a transverse electric field [6,7]. Another technique, that requires a MCP for low intensity beams, exploits the secondary emission of electrons from wires (tungsten) and/or thin foils (carbon or aluminium) when hit by energetic particles. A gas detector with micro strip readout electrode has shown to be rather stable, reliable and interactive; though it is non-destructive and very sensitive, it has the disadvantage of being a gas device along a beam line, involving the well known risks.

Beam identification

In order to perform the beam identification after acceleration and recognize the present nuclear species, we have developed a device named HEBI (High Energy Beam Identifier), based on a high resolution silicon telescope, that can revolve around a target. It is composed of two parallel detectors, the DE and the thicker E. The capability of this system to identify the nuclei with high efficiency, allows to determine the nuclear species present in the beam.

The silicon telescope can be accurately positioned around a target (typically gold), placed along the beam line, in order to intercept the scattered ions. The angle where the telescope must be placed is chosen as a function of the expected ions and of their energy, in order to have an intensity not larger than 10⁴ particles per second on the telescope, to prevent fast detector damage. This angle can be estimated by means of the well known Rutherford cross section. However this angle must be lower than the grazing angle, since it is necessary to avoid the predominance of reaction products. For very low beam intensity (I_beam < 10⁴ pps), the gold target can be removed and the telescope put at 0° degrees along the beam axis.

In order to study the discrimination efficiency, a test has been done with a ¹⁶O beam at 10⁹ pps intensity. An alpha source was placed close to the telescope, and has been used to perform the calibration procedure of the system and for the energy resolution measurement. The silicon telescope was composed of a 20 mm thick DE and a 400 mm thick E layers. It was placed at a distance of 25 cm from a ¹⁹⁶Au target, at an angle of 10° with respect to the beam axis, so that the rate of ions hitting the detector was about 10² pps. The calibration procedure was performed by exploiting the three energetic peaks of the alpha source and the elastic scattering peak of the oxygen.

The data acquired during this test allowed to build the DE-E scatter plot shown in fig.9. Each peak acquired by DE detector has been fitted by a gaussian function in order to estimate s, and then we found that plotting s versus the root square of the centroid energy , the experimental points are well fitted by the following linear function:

s = 0.0835× + 0.018 (MeV)

We can reasonably suppose that in the range of energy and mass of the beams under investigation, the s interpolated by the previous equation is reliable enough, thus we use it to calculate the error bands for the nuclear species to be produced with EXCYT and their respective isobaric contaminants. For this purpose we have taken into consideration three elements: ¹¹Be, ¹⁷F and ¹⁸F. For each of these and their isobaric contaminants, we calculated the energy loss in the DE detector, in order to build the relative DE-E plot, with the respective error bands (+ s) . In fig. 10 the plot containing the calculated curves for the ¹⁷F and contaminants is shown. We have calculated the probability that a beam particle is misidentified with a contaminant, and we found that it is always below 10 ^-10.

Fig. 9. Calibrated bands (±s) superimposed to the experimental data taken with HEBI. From the ¹⁶O+¹⁹⁶Au reaction we get mainly elastic scattering, plus many alphas and some C product.

Fig. 10. Discrimination plot (±s) for ¹⁷F. The two main contaminants are shown.

References

[1] G. Ciavola et al., Nucl. Instr. and Meth. B126 (1997) 258-261;

[2] S.Cappello, L.Cosentino, P.Finocchiaro, Nucl. Instr. & Meth. A479(2002)243;

[3] L.Cosentino, P.Finocchiaro, IEEE Trans. Nucl. Sci. Vol.48, No.4, (2001)1132;

[4] P. Finocchiaro et al., Nucl. Instr. and Meth. A419 (1998) 83-90;

[5] P. Finocchiaro et al., Nucl. Instr. and Meth. A437 (1999) 552-556;

[6] P. Finocchiaro, DIPAC 97, 3rd European Workshop on Beam Diagnostic and Instr. for Particle Accel., Lab. Naz: Frascati, October 14th, 1997;

[7] P. Finocchiaro, CAARI 98, 15th Internat. Conf. on the Appl. of Accel. In Research and Industry, Univ. of North Texas Denton, November 4-7, 1998.