— journal of September 2015
physics pp. 497–504
Nuclear fission as a tool to contrast the contraband of special nuclear material
VIESTI GIUSEPPE1, CESTER DAVIDE1, NEBBIA GIANCARLO2,∗, STEVANATO LUCA1, NERI FRANCESCO3, PETRUCCI STEFANO3, SELMI SIMONE3and TINTORI CARLO3
1Dipartimento di Fisica ed Astronomia, Università di Padova, Via Marzolo 8, Padova I-35131, Italy
2INFN Sezione di Padova, Via Marzolo 8, Padova I-35131, Italy
3CAEN S.p.A., Via Vetraia 11, I-55049, Viareggio (LU), Italy
∗Corresponding author. E-mail: giancarlo.nebbia@pd.infn.it
DOI:10.1007/s12043-015-1061-1; ePublication:25 August 2015
Abstract. An integrated mobile system for port security is presented. The system was designed to perform passive measurements of neutrons andγ-rays to search and identify radioactive and special nuclear materials as well as for the active investigations by using the tagged neutron inspection technique of suspect dangerous materials. The discrimination between difficult-to-detect uranium samples and highZmaterials as lead was specifically studied. The system has been employed in laboratory detection tests and in a seaport field test.
Keywords. Associated particle technique; fast neutron inspection; detection of special nuclear material.
PACS Nos 28.20.−V; 29.30.Kv; 89.20.Bb; 89.40.−Cc
1. Introduction
The SLIMPORT project, financed by the Italian Ministry for the Economic Development (MISE), was dedicated to the development of an integrated package of tools forming a complete security system to monitor transport of persons and merchandise in seaports.
In this framework, a mobile inspection station (called SMANDRA, the Italian acronym stands for Sistema Mobile per Analisi Non Distruttive e RAdiometriche i.e., Mobile Sys- tem for Non-Destructive Analysis and Radiometric Measurements) has been developed.
The aim of SMANDRA is to search and identify sources of ionizing radiation by pas- sive measurements ofγ-rays and neutrons or to identify dangerous and/or illegal mate- rials inside volumes tagged as ‘suspect’ by previous conventional surveys such as
X-ray scans. Thus the SMANDRA system is a second line active inspection tool. The SMANDRA system consists of two units:
(1) A ‘passive unit’ consisting of twoγ-ray detectors (5′′×5′′ NaI(Tl) and 2′′×2′′
LaBr(Ce) scintillation detectors) and two neutron counters (5′′×2′′liquid scintil- lator and3He proportional counter for fast/slow neutron measurements). This unit hosts batteries, power supplies, front-end electronics and CPU.
(2) An ‘active unit’ consisting of a portable sealed neutron generator type TPA-17 from EADS-SODERN.
The first unit can be used in standalone mode as a high-efficiency spectroscopic radiometer for the detection of ionizing radiation, to search and identify radioactive mate- rial as well as special nuclear material (SNM). It can also be used as detector package connected to the ‘active unit’ for the interrogation of voxels inside a load by means of the tagged neutron inspection system (TNIS) technique [1].
The SMANDRA system has been fully described in [2] and only a short description is provided here. Laboratory results obtained so far are also summarized [3,4] discussing in detail the active detection of SNM. Finally, results from the recent field demonstration are presented.
2. The SMANDRA system
The complete SMANDRA system during laboratory tests with SNM is shown in figure 1.
The dual use of SMANDRA system (active and passive interrogations) sets stringent requirements:
(1) Low background, high-efficiency detectors forγ-rays and neutrons, with the need to discriminate the two components of radiation in the passive mode use.
(2) High count rate capability detectors to be operated in coincidence with the associated particle counter hosted in the neutron generator.
For γ-ray detection, photon spectroscopy is performed by using both the high- resolution LaBr(Ce) detector and the high-efficiency NaI(Tl) scintillator. The LaBr(Ce) detector offers the ultimate energy resolution for scintillators but the one used in SMANDRA has a limited volume compared to other scintillation materials. Moreover, LaBr(Ce) suffers from the internal activity that sets limits in the capability of detecting weak sources [3]. Consequently, the NaI(Tl) scintillator was selected to be used as a high- efficiency device for the detection and identification of weak γ sources with a simple decay scheme, when the energy resolution is not required to discriminateγ transitions with similar energies. As for the neutron detectors, the3He proportional counter with a polyethylene moderator is a typical choice for such systems operated in passive mode.
This counter provides the information about the total neutron yield without the possibility of discriminating fast from thermal neutrons. However, the direct detection of fast neu- trons both in passive and active modes is an important task that justifies the use of liquid organic scintillator (a 5′′×2′′cell). It is also worth mentioning that the time resolution of liquid scintillators is very important in performing active interrogations.
Figure 1. The SMANDRA system during active interrogation of a highly enriched uranium sample placed in front of the detector unit. The box containing the EADS- SODERN TPA-17 neutron generator is shown on the left.
In the SMANDRA system, both types of measurements (passive and active) are man- aged by a simple CAEN VME electronic front-end based on fast digitizers. The front end makes use of a battery-operated VME minicrate (four slots) with a Bridge USB V1718.
The minicrate hosts a HV system (V6533 Programmable HV Power Supply (six Ch., 4 kV, 3 mA, 9 W) and a V1720 eight channel 12 bit 250 MS/s Digitizer.
Inside the V1720, digital pulse processing (DPP) algorithms are implemented by using field programmable gate array (FPGA), providing online, the time stamp for each event, the complete integration of the signal, a partial integration of the signal used for pulse shape discrimination in the liquid scintillator and the possibility of storing a selected part of the digitized signal. The latter feature is required to reconstruct off-line coincidences and for the time measurements in active mode.
Intense laboratory work has been carried out to characterize the detector performances with the VME front-end by comparing with the data obtained from conventional NIM electronics read-out. In particular, the NaI(Tl) and LaBr(Ce) detectors have been char- acterized in terms of energy resolution and pulse amplitude stability as a function of counting rate. Also LaBr(Ce) with digital signal processing has been operated up to a rate of 340 kHz with excellent results [3].
The neutron-gamma discrimination is also performed online by the FPGA which pro- vides both the total integration of the liquid scintillator signal (total light) and the integra- tion of the prompt part of the distribution (prompt light). The ratio between the delayed light (obtained by the difference between the total and the prompt ones) and the total light is used to perform online pulse shape discrimination as a function of the total light [2].
In active interrogations the associated particle detector signal is also processed in the V1720 card. The α-particles emitted in the 3H(2H,4He)n neutron source reaction are indeed detected in a fast YAP(Ce) scintillation detector embedded inside the neutron generator, coupled to an external HAMAMATSU R1450 PMT.
According to the results of laboratory-active interrogations performed so far, the asso- ciated particle detector covers a fraction of solid angle of about 10−3such that a rate of 10 kHz characterizes the operation of the neutron generator at a total flux of 107neutrons/s.
Under this condition, the spot of the tagged neutron beam at the object position (located 30 cm from the ‘passive unit’ front face) has been measured to be about 15 cm (FWHM), depending on the acceptance of the YAP(Ce) detector. In the active mode operation we directly stored all the event singles processed by the V1720 card running at a typical total rate of about 50 kHz recording the interesting part of the digitized signals. Offline software analyses the event files reconstructing the coincidence events and the time cor- relation between detectors. The time resolution depends on the way of handling the data.
The best results have been obtained using a ‘digital constant fraction discriminator’ [2]
that allows one to obtain time resolutions of the order of aboutδt = 1 ns (FWHM) for the LaBr(Ce) andδt =5 ns (FWHM) for the NaI(Tl) detectors with the thresholds set at 500 keV.
It is worth mentioning that a 5 ns time resolution reflects a depth resolution of about 25 cm for the inspected voxel for 14 MeV tagged neutrons.
3. Detection and identification of radioactive sources
Laboratory tests have been carried out to verify the possibility of detecting the presence of radioactive material (γ-ray or neutron sources) and identifying the type of sources.
As a guidance, the IEC 62327 standard for hand-held instruments for the detection and identification of radionuclides has been considered. A 3 s time lapse has been selected to verify the presence of alarms in NaI(Tl) forγ-rays, whereas a 10 s cycle is used for neutrons. SMANDRA detects a 0.4 MBq60Co source at 270 cm from the front face of the detector (with an equivalent dose of 20 nSv/h) and 0.4 MBq241Am at 80 cm from the front face of the detector (with an equivalent dose of 2.5 nSv/h) with PD=90% at CF=95%. This result needs to be compared with the IEC62372 requirement of detection for a source that produces 500 nSv/h at the front face of the detector. After the alarm, the identification ofγ-source requires a measurement of 1 min.
For neutrons, SMANDRA, after a proper energy windowing, detects in 10 s, the weak
252Cf source placed at 140 cm from the detector surface with PD>90% at CF=95%, demonstrating a sensitivity about 60 times larger than required by the IEC62372.
4. Detection and identification of special nuclear material
A campaign dedicated to the detection of SNM has been carried out at the PERLA Lab- oratory of the Joint Research Center of the European Commission in Ispra using several Pu and U samples having different enrichments and weights [4]. SMANDRA was used in passive as well in active modes. To summarize, the Pu samples were easily identified by their neutron emission and characteristicγ-ray signatures fully exploited in the LaBr(Ce) detector.
The detection of U samples appeared difficult for which the neutron yield is quite low and the characteristicγ-ray signature is also low in energy (and then easily masked or attenuated by shielding). This is demonstrated in figure 2 where the effect of attenuation due to lead shielding is detailed for the characteristicγ-rays emitted from a 1 g source of weapon-grade plutonium (93%239Pu) and uranium (93%235Pu). In case of WGPu, the
239Pu and 241Am transitions (Eγ = 414 and 662 keV) still have a yield of about 100 Hz after 2.5 cm of shielding making the detection possible at close contact.
Consequently, it appeared interesting to study the detection of U samples in active interrogations.
Several chemicals and uranium samples have been bombarded with typical measuring time of 10 min at a neutron total flux of 107neutron/s. First, it is verified that inorganic or iron-based materials are easily identified by their well-known coincidentγ-ray spectrum whereas the U samples cannot be discriminated from other heavy metals such as Pb that exhibit featureless coincidentγ-ray spectra.
It has been verified that the discrimination between U samples and other materials can be achieved by analysing the correlation between SMANDRA detectors. In particular, it was observed that when the liquid scintillator is used to separateγ-rays from neu- trons, a good U discrimination is achieved on plotting the ratio between the triple (liquid scintillator–NaI(Tl)–YAP(Ce)) and the double (liquid scintillator–YAP(Ce)) coincidences inγ-ray and neutron events [4]. Typical results are shown in figure 3. In other words, as the NaI(Tl) mainly detectsγ-rays because of the difference in its intrinsic efficiency for photons and neutrons, we are plotting the probability of multiple γ events vs. the γ–neutron coincidences, and this plot clearly separates U from other samples, including
Figure 2. Gamma-ray yield from 1 g weapon-grade plutonium and uranium sam- ples as a function of the thickness of lead shield:241Am,Eγ = 662 keV (diamond);
239Pu, Eγ = 414 keV (square); 235U, Eγ = 186 keV (yellow triangle); 238U, Eγ =1001 keV (green triangle).
Figure 3. Correlation between the triple (α-particle–liquid scintillator–NaI(Tl) detec- tors) and double (α-particle–liquid scintillator detectors) coincidences for neutrons andγ-rays identified in the liquid scintillator. The square refers to iron, the diamond to lead, the cross to organic and the triangles to uranium (full triangle denotes 2.5 kg sample with 4.4% enrichment on235U, empty triangle denotes 2.5 kg sample with enrichment on235U).
lead, as demonstrated in [4]. This was due to the presence of neutron-induced fission in U samples. More recently, we have also studied the effect of shielding on this type of dis- crimination by simply using a252Cf source to produce fission events. Results are reported in figure 4 relative to the unshielded source. It seems that using 1 or 2 cm lead shielding immediately inhibits the detection of γ-rays in the NaI(Tl) detectors thus lowering the triple/double ratio. This means that a shielded fission source will be easily confused with other materials.
Figure 4. Dependence of the triples/doubles ratio forγ-rays and neutrons as a func- tion of lead shield thickness when NaI(Tl) detector is used to build the triple events (squares). The triangles refer to a system in which NaI(Tl) is replaced by a liquid scintillator where only neutrons are selected. The data without shield are normalized to (1, 1) point whereas the other data points refer to 1 and 2 cm lead shields.
A second test was performed by replacing the NaI(Tl) detector with a second liquid scintillator in which only neutrons are selected by the software. In this case, as the neu- trons are scarcely attenuated by lead, the triple/double ratio of a fissile source remains close to that of an unshielded source. This evidence will guide us to design a new version of SMANDRA ‘passive unit’ that is able to distinguish fission sources in the presence of different shielding.
5. Field demonstration
The ‘passive unit’ of the SMANDRA system has been employed recently in a field demonstration at La Spezia seaport (Italy) together with other participants to the task SlimChek of the SLIMPORT project, as documented in figure 5.
The demonstration was directed to the National Firefighter Corp and was structured in the following way:
(1) The SMANDRA system was used to determine the position of a weak radioac- tive source (about 20 kBq) located inside a shipping container and to identify the radioactive material.
(2) After that a remote-controlled forklift entered the container to remove some pellets of materials around the source position.
(3) Finally, a remote-controlled robotic arm entered the container for catching and transporting the source on a safety dump located outside the container.
This demonstration was very successful.
Figure 5. Pictures of the seaport demonstration. The SMANDRA system showing the robotic arm (a) and searching for radioactive source (b).
6. Summary and conclusions
The mobile SMANDRA inspection system has been tested in laboratory conditions for two distinct tasks: as a high sensitivity passive spectroscopic system and as a com- plete inspection system using tagged neutrons. Uranium samples are discriminated from non-fissile heavy elements by taking advantage of the large fission cross-section that sig- nificantly increases the possibility of detection for neutron–γ-ray or neutron–neutron coincidences. Results obtained so far demonstrate the good capability of the present prototype and will guide us in preparing a more advanced version of the system.
References
[1] S Pesente, G Nebbia, M Lunardon, G Viesti, S Blagus, K Nad, D Sudac, V Valkovic, I Lefesvre and M J Lopez-Jimenez,Nucl. Instrum. Methods B241, 743 (2005)
[2] D Cester, D Fabris, M Lunardon, S Moretto, G Nebbia, S Pesente, L Stevanato, G Viesti, F Neri, S Petrucci, S Selmi and C Tintori,IEEE Proc. ANIMMA, ISBN number 978-1-4577-0926-5, 6172933 (2011)
[3] L Stevanato, D Cester, G Nebbia, G Viesti, F Neri, S Petrucci, S Selmi and C Tintori,Nucl.
Instrum. Methods A678, 83 (2012)
[4] D Cester, G Nebbia, L Stevanato, G Viesti, F Neri, S Petrucci, S Selmi, C Tintori, P Peerani and A Tomanin,Nucl. Instrum. Methods A663, 55 (2012)
