A high-resolution neutron spectra unfolding method using the Genetic Algorithm technique

The Bonner sphere spectrometers (BSS) are commonly used to determine the neutron spectra within various nuclear facilities. Sophisticated mathematical tools are used to unfold the neutron energy distribution from the output data of the BSS. This paper highlights a novel high-resolution neutron spectra-unfolding method using the Genetic Algorithm (GA) technique. The GA imitates the biological evolution process prevailing in the nature to solve complex optimisation problems. The GA method was utilised to evaluate the neutron energy distribution, average energy, fluence and equivalent dose rates at important work places of a DIDO class research reactor and a high-energy superconducting heavy ion cyclotron. The spectrometer was calibrated with a ²⁴¹Am/Be (α,n) neutron standard source. The results of the GA method agreed satisfactorily with the results obtained by using the well-known BUNKI neutron spectra unfolding code.     

Evaluation of individual monitoring in mixed neutron/photon fields: mid-term results from the EVIDOS project

EVIDOS is an EC sponsored project that aims at an evaluation and improvement of radiation protection dosimetry in mixed neutron/photon fields. This is performed through spectrometric and dosimetric investigations during different measurement campaigns in representative workplaces of the nuclear industry. The performance of routine and, in particular, novel personal dosemeters and survey instruments is tested in selected workplace fields. Reference values for the dose equivalent quantities, H(*)(10) and H(p)(10) and the effective dose E, are determined using different spectrometers that provide the energy distribution of the neutron fluence and using newly developed devices that determine the energy and directional distribution of the neutron fluence. The EVIDOS project has passed the mid-term, and three measurement campaigns have been performed. This paper will give an overview and some new results from the third campaign that was held in Mol (Belgium), around the research reactor VENUS and in the MOX producing plant of Belgonucléaire.     

Characterisation of mixed neutron photon workplace fields at nuclear facilities by spectrometry (energy and direction) within the EVIDOS project

Within the EC project EVIDOS, 17 different mixed neutron-photon workplace fields at nuclear facilities (boiling water reactor, pressurised water reactor, research reactor, fuel processing, storage of spent fuel) were characterised using conventional Bonner sphere spectrometry and newly developed direction spectrometers. The results of the analysis, using Bayesian parameter estimation methods and different unfolding codes, some of them especially adapted to simultaneously unfold energy and direction distributions of the neutron fluence, showed that neutron spectra differed strongly at the different places, both in energy and direction distribution. The implication of the results for the determination of reference values for radiation protection quantities (ambient dose equivalent, personal dose equivalent and effective dose) and the related uncertainties are discussed.     

Measurement of the spectral fluence rate of reference neutron sources with a liquid scintillation detector

Reference neutron sources such as (241)AmBe(alpha,n) and (252)Cf are commonly used to calibrate neutron detectors for radiation protection purposes. The calibration factors of these detectors depend on the spectral distribution of the neutron fluence from the source. Differences between the spectral fluence of the neutron source and the ISO-recommended reference spectra might be caused by the properties of the individual source. The spectral neutron fluence rates of different reference neutron sources used at PTB were measured with a liquid scintillation detector (NE213), using maximum entropy unfolding and a new, experimentally determined detector response matrix. The detector response matrix was determined by means of the time-of-flight technique at a pulsed neutron source with a broad energy distribution realised at the PTB accelerator facility. The results of the measurements of the reference sources are compared with the ISO-recommended reference spectra. For the PTB (241)AmBe(alpha,n) reference source, the spectral neutron fluence was determined by means of a high-resolution (3)He semiconductor sandwich spectrometer in 1982. These measurements were the basis for the ISO recommendations. The current measurements confirm the high-energy part (E(n) > 2 MeV) of this spectrum and demonstrate the suitability of this new method for high-resolution spectrometry of broad neutron spectra.     

A directional dose equivalent monitor for neutrons

A directional dose equivalent monitor is introduced which consists of a 30 cm diameter spherical phantom hosting a superheated drop detector embedded at a depth of 10 mm. The device relies on the similarity between the fluence response of neutron superheated drop detectors based on halocarbon-12 and the quality-factor-weighted kerma factor. This implies that these detectors can be used for in-phantom dosimetry and provide a direct reading of dose equivalent at depth. The directional dose equivalent monitor was characterised experimentally with fast neutron calibrations and numerically with Monte Carlo simulations. The fluence response was determined at angles of 0, 45, 90, 135 and 180 degrees for thermal to 20 MeV neutrons. The response of the device is closely proportional to the fluence-to-directional dose equivalent conversion coefficient, h'phi (10; alpha, E). Therefore, our monitor is suitable for a direct measurement of neutron directional dose equivalent, H'(10), regardless of angle and energy distribution of the neutron fluence.     

A wide-range direction neutron spectrometer

A new device is presented which has been developed for measuring the energy and direction of distribution of neutron fluence in fields of broad energy spectra (thermal to 100 MeV) and with a high background of photon, electron and muon radiation. The device was tested in reference fields with different energy and direction distributions of neutron fluence. The direction-integrated fluence spectra agree fairly well with reference spectra. In all cases, the ambient and personal dose equivalent values calculated from measured direction-differential spectra are within 35% of the reference values. Independent measurements of the directional dose equivalent were performed with a directional dose equivalent monitor based on superheated drop detectors.     

Characterisation of the IRSN graphite moderated Americium-Beryllium neutron field

The SIGMA facility was set up at IRSN to provide thermal neutrons for metrology and dosimetry purposes. SIGMA consists of six Am-Be radioactive sources located in a 1.5 x 1.5 x 1.5 m3 graphite moderator block. The neutron field at the calibration position, situated at 50 cm from the west surface of the assembly was characterised experimentally and by Monte Carlo calculations. The thermal neutron fluence was determined by the activation of gold foils; the neutron fluence energy distribution above 240 keV was measured with proton recoil spectrometers and the neutron fluence energy distribution from thermal energies to 20 MeV was measured with a Bonner spheres spectrometer. A Monte Carlo simulation of the SIGMA assembly was undertaken using the MCNP4C code, and the calculated neutron fluence energy distribution was compared with the measurements. As a whole, the experimental data and the MCNP calculation are in a good agreement.     

Monte Carlo simulation of the IRSN CANEL/T400 realistic mixed neutron-photon radiation field

The calibration of dosemeters and spectrometers in realistic neutron fields simulating those encountered at workplaces is of high necessity to provide true and reliable dosimetric information to the exposed nuclear workers. The CANEL assembly was set-up at IRSN to produce such neutron fields. It comprises a depleted uranium shell, to produce fission neutrons, then iron and water to moderate them and a polyethylene duct. The new presented CANEL facility is used with 3.3 MeV neutrons. Calculations were performed with the MCNP4C code to characterise this mixed neutron-photon expanded radiation field at the position where calibrations are usually performed. The neutron fluence energy and the direction distributions were calculated and the operational quantities were derived from these distributions. The photon fluence and corresponding ambient dose equivalent were also estimated. Comparison with experimental results showed an overall good agreement.     

Bonner sphere neutron spectrometry at nuclear workplaces in the framework of the EVIDOS project

The Institute for Radiological protection and Nuclear Safety was engaged in the EC funded EVIDOS project to provide reference spectrometry data using its Bonner sphere system. The data were processed by means of two unfolding codes, NUBAY and GRAVEL, both provided by the Physikalisch-Technische Bundesanstalt. The NUBAY program, based on Bayesian parameter estimation methods, assumes a parameterised spectrum and provides posterior probability distributions for the parameters. The code GRAVEL, an iterative algorithm based on SAND-II, was used with various default spectra, among them the NUBAY solution. The BS measurements were used to establish the neutron fluence energy distributions and reference values for the neutron ambient dose equivalent. As this quantity depends strongly on the high energy neutrons, a sensitivity analysis was done by unfolding the BS data with GRAVEL using the NUBAY solution spectrum as default with various changes in the parameters of the high energy peak. This new method of analysing Bonner sphere data allowed the determination of reliable neutron spectra, as well as a very good estimate of the corresponding integral quantities with small associated uncertainties.     

A simple method for the determination of the neutron dose in a phantom

A method based on a combination of physical integration and activation threshold detectors was developed to determine the volume averaged dose equivalent rates produced by 14.1 MeV incident neutrons in a water filled phantom. To obtain the spectral fluence of neutrons in phantoms activation threshold detector measurements and a least-squares unfolding code (LSQ) were used. The physical integration was carried out by stirring the phantom solution after irradiation. The method is suitable also to determine the energy averaged conversion factor between the maximum dose equivalent and the primary fast neutron fluence measured on the surface of the phantom. The method proposed can be applied for any kind of phantom geometry.     

A compact high-energy neutron spectrometer

A compact liquid organic neutron spectrometer based on a single NE213 liquid scintillator (5 cm diameter x 5 cm) is described. The spectrometer is designed to measure neutron fluence spectra over the energy range 2-200 MeV and is suitable for use in neutron fields having any type of time structure. Neutron fluence spectra are obtained from measurements of two-parameter distributions (counts versus pulse-height and pulse shape) using the Bayesian unfolding code MAXED. Calibration and test measurements made using a pulsed neutron beam with a continuous energy spectrum are described and the application of the spectrometer to radiation dose measurements is discussed.     

Neutron spectrometry for radiation protection purposes

Determination of the dose equivalent is required for radiation protection purposes, however such a determination is quite difficult for neutron radiation. In order to perform accurate dosimetric determinations, it is advantageous to acquire information about the neutron fluence spectrum in the workplace as well as the reference radiations used to calibrate dosimetric instruments. This information can then be used to select the appropriate dosimetric instrument, the optimum calibration condition or to establish correction factors that account for the differences in calibration and workplace conditions. For quite some time, neutron spectrometry has been used for these purposes. A brief review of the applications of spectrometers in radiation protection and some recommendations for further development are given here.     

A fundamental study of neutron spectra unfolding based on the maximum likelihood method

A new unfolding method, which treats the Poisson statistics of neutron detection correctly, is developed on the basis of the maximum likelihood method. The likelihood equations derived here are nonlinear with respect to neutron spectrum, but solved with the aid of the “singular value decomposition” of the response matrix. The solution thus obtained is unbiased, preserves total counts, and reduces to the conventional least-squares one if approximated. In the course of numerical calculation with the use of idealized rectangular response functions, it has been found that the choice of the number of pulse-height bins is very important in diminishing the errors of the resultant spectra for a given number of energy groups. To evaluate a variety of unfolded results, a new quantity named “error magnification rate” is proposed whose asymptotic value indicates an intrinsic limitation spectrometers requiring an unfolding procedure.     

Neutron spectrometry in mixed fields: superheated drop (bubble) detectors

The BINS neutron threshold spectrometer permits the analysis of the main features of a neutron field for radiation protection purposes. The system offers a virtually complete photon discrimination and nested threshold responses to neutrons, which allow the use of very effective 'few-channel' unfolding procedures. To date, the practical operating energy range of a BINS is 0.1-10 MeV, over which a resolving power of 20-30% can be expected when the deconvolution is performed without explicit pre-information. Spectrum unfolding results in relatively high uncertainties on the differential fluence distributions, but due to negative correlations in adjacent energy groups the uncertainties on integral quantities such as dose equivalent are small and of the order of 5% to 10%, similar to the results of other active spectrometers. In comparison with most radiation detectors, the BINS is an extremely slow system due to the intrinsic duration of a bubble pulse and to the time associated with pulse analysis. For example, the maximum sustainable fluence rate of 1 MeV neutrons is about 10(4) cm(-2) s(-1), which is low for many neutron physics experiments. However, this rate corresponds to an ambient dose equivalent rate of about 1 mSv h(-1), making the active device adequate for radiation protection applications in the workplaces described in Section 1. There are ample margins for improvement of the spectrometer. In particular, in the low-energy region a thermal-epithermal neutron group may be added by using chlorine-bearing emulsions stabilised at suitable temperatures. In fact, the latest version of the system achieves this goal by using a single superheated emulsion of dichlorotetrafluoroethane (R-114) operated at temperatures up to 55 degrees C. This extends the range of the spectrometer and at the same time removes the undue enhancement of the UNFANA output in the low energy region. Above 10 MeV, the resolution can be improved by adding more thresholds, e.g. by starting from a lower initial temperature and using finer temperature increments. Based on neutron kinematics, the theoretical upper energy threshold which can be generated with superheated emulsions is greater than 100 MeV. However, this would most likely require refrigerating the detectors, while the current simpler approach is to operate the detectors at incremental temperature steps starting from the ambient temperature. A range that should be easily achieved in practice is from thermal energies to 20 MeV.     

Advantages of passive detectors for the determination of the cosmic ray induced neutron environment

Due to the pronounced energy dependence of the neutron quality factor, accurate assessment of the biologically relevant dose requires knowledge of the spectral neutron fluence rate. Bonner sphere spectrometers (BSSs) are the only instruments which provide a sufficient response over practically the whole energy range of the cosmic ray induced neutron component. Measurements in a 62 MeV proton beam at Paul Scherrer Institute, Switzerland, and in the CERN-EU high-energy reference field led to the assumption that conventional active devices for the detection of thermal neutrons inside the BSS, e.g. 6Lil(Eu) scintillators, also respond to charged particles when used in high-energy mixed radiation fields. The effects of these particles cannot be suppressed by amplitude discrimination and are subsequently misinterpreted as neutron radiation. In contrast, paired TLD-600 and TLD-700 thermoluminescence dosemeters allow the determination of a net thermal neutron signal.     

Investigation of combined unfolding of neutron spectra using the UMG unfolding codes

An investigation of the simultaneous unfolding of data from neutron spectrometers using the UMG codes MAXED and GRAVEL has been performed. This approach involves combining the data from the spectrometers before unfolding, thereby performing a single combined unfolding of all the data to yield a final combined spectrum. The study used measured data from three proton recoil counters and also Bonner sphere and proton recoil counter responses calculated from their response functions. In each case, the spectrum derived from combined unfolding is compared with either the spectrum obtained from merging the independently unfolded spectra or the spectrum used to calculate the responses. The advantages and disadvantages of this technique are discussed.     

A telescope-design directional neutron spectrometer

A directional spectrometer that uses a superheated emulsion of dichlorotetrafluoroethane at the centre of a 30 cm diameter moderating-sphere of nylon-6. The system has a telescope-design wherein the detector views a narrow solid angle of about 1/6 steradians. The hydrogenous sphere effectively attenuates laterally incident neutrons, thus providing a strong angular dependence of the response. The central detector is sequentially operated at seven temperatures between 25 and 55 degrees C in order to generate a matrix of nested response functions suitable for few-channel spectrometry. The response matrix of the system has been determined by calibrations with monoenergetic neutrons and by Monte Carlo neutron transport calculations. The double-differential unfolding method developed for this system applies the principle of maximum entropy and allows for the rigorous use of all a priori information. The spectrometer is intended for use in the mixed neutron/photon fields encountered in the nuclear power industry, being suitable for spatially distributed radiation sources with maximum neutron energies up to 10 MeV.     

Performance of the UAB and the INFN-LNF Bonner sphere spectrometers in quasi monoenergetic neutron fields

In the framework of collaboration between the Universidad Autónoma de Barcelona and the INFN Frascati National Laboratories, an experimental Bonner Sphere neutron spectrometry exercise has been performed in the 2.5 MeV and 14.2 MeV quasi monoenergetic neutron beams of the ENEA Fast Neutron Generator. The neutron spectra at given distances from the accelerator target have been determined, taking advantage of the new unfolding FRUIT code, recently developed by the LNF group. The results show a good coherence between the two spectrometers, and between the measured and simulated data.     

Results of field trials using the NPL simulated reactor neutron field facility

The NPL simulated reactor neutron field facility provides neutron spectra similar to those found in the environs of UK gas-cooled reactors. Neutrons are generated by irradiating a thick lithium-alloy target with monoenergetic protons between 2.5 and 3.5 MeV (depending on the desired spectrum), and then moderated by a 40-cm diameter sphere of heavy water. This represents an extremely soft workplace field, with a mean neutron energy of 25 keV and, more significantly, a mean fluence to ambient dose equivalent conversion coefficient of the order of 20 pSv cm(2), approximately 20 times lower than those of the ISO standard calibration sources (252)Cf and (241)Am-Be. Results of field trials are presented, including readings from neutron spectrometers, personal dosimeters (active and passive) and neutron area survey meters, and issues with beam monitoring are discussed.     

Measurement of neutron dose with an organic liquid scintillator coupled with a spectrum weight function

A dose evaluation method for neutrons in the energy range of a few MeV to 100 MeV has been developed using a spectrum weight function (G-function), which is applied to an organic liquid scintillator of 12.7 cm in diameter and 12.7 cm in length. The G-function that converts the pulse height spectrum of the scintillator into the ambient dose equivalent, H*(10), was calculated by an unfolding method using successive approximation of the response function of the scintillator and the ambient dose equivalent per unit neutron fluence (H*(10) conversion coefficients) of ICRP 74. To verify the response function of the scintillator and the value of H*(10) evaluated by the G-function. pulse height spectra of the scintillator were measured in some different neutron fields, which have continuous energy, monoenergetic and quasi-monoenergetic spectra. Values of H*(10) estimated using the G-function and pulse height spectra of the scintillator were compared with those calculated using neutron energy spectra. These doses agreed with each other. From the results, it was concluded that H*(10) can be evaluated directly from the pulse height spectrum of the scintillator by applying the G-function proposed in this study.