In-situ element analysis from gamma-ray and neutron spectra using a pulsed-neutron source Maleka, Peane Peter

Transcription

1 University of Groningen In-situ element analysis from gamma-ray and neutron spectra using a pulsed-neutron source Maleka, Peane Peter IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2010 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Maleka, P. P. (2010). In-situ element analysis from gamma-ray and neutron spectra using a pulsed-neutron source Groningen: s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 CHAPTER 5 FAST-NEUTRON SPECTRA ANALYSIS 5.1 Introduction Before 1930, the only penetrating radiation known was gamma radiation (Allen, 1960). When Von Bothe and Becker reported that a very penetrating radiation was obtained when light elements such as boron or beryllium were bombarded with alpha particles, they assumed this radiation also to be gamma-rays with energies in the range of 7-15 MeV, (Von Bothe and Becker 1930; Allen, 1960). But their conclusion was challenged by Curie and Joliot, by concluding that the gamma-rays had energies of 50 MeV instead (Chadwick, 1932; Allen, 1960). In 1932, Chadwick questioned the gamma-ray hypothesis because the cross sections involved in the reaction and very high energies required for the gamma-rays were improbable (Chadwick, 1932). He showed that assuming that the radiation consists of neutral particles with about the proton mass, these difficulties could be removed. The discovery of the neutron rested on the measurements of neutron energy made by scattering neutrons on hydrogen (Chadwick, 1932) or nitrogen (Feather, 1932), and measuring the resulting recoiling energies. Since then neutron spectroscopy has contributed significantly to the development of nuclear physics and has also become an important tool in other fields: nuclear technology, radiotherapy and radiation protection (Brooks and Klein, 2002). One of the methods in neutron-spectroscopy is based on measuring the energies of charged particles released in neutron-induced reactions (Brooks and Klein, 2002). Neutron reactions with material depend mainly on the neutron energy and the nuclides in the material. By monitoring these neutron interactions, it is possible to determine the nuclides and hence their corresponding chemical elements. In section 2.6 we discussed the process that leads to neutron detection. For the NuPulse instrument, the D-T neutron generator (see section 3.2) and neutron detectors (see section 3.7) are combined in experimental tests. In this chapter, we focus mainly on the configuration that involves fast-neutron detectors. An organic stilbene crystal (5.5 cm in diameter and length) is used because of its capability to distinguish gamma-rays and neutrons based on their pulse shape. The D-T neutron generator will emit neutrons with energies of about 14 MeV (Knoll, 2000) and their reactions with surrounding materials are investigated. Originally in the NuPulse project, fast-neutron spectra were merely regarded as a mode to monitor the neutron flux produced by the D-T source by integrating the neutron-energy spectrum. In the course of this thesis work, fast-neutron spectra were further investigated as a tool to extract the chemical composition of the environment surrounding the NuPulse instrument. 5.2 Stilbene detector calibration The organic stilbene detector systems were assembled and calibrated by the NuPulse subcontractor to SELOR, SIOL from Kharkov, Ukraine (see also section 3.7.1). For energy calibrations of the stilbene detectors, reference gamma-ray sources are used and their (gamma-rays) energies are subsequently matched with the neutron energies. Currently reference sources emitting neutrons with specific energy are not 103

3 Number of pulses FAST-NEUTRON SPECTRA ANALYSIS readily available. The stilbene scintillators distinguish fast neutrons from gamma-rays on the basis of their difference in decay times of recoil protons (neutrons) and Compton electrons (gamma-rays), (Brooks, 1959). All energy calibrations for the NuPulse fast-neutron detectors were carried out by SIOL to control/reject gamma-rays signals. Thereafter, it was not possible for other partners to redo/check the energy calibration using the gamma-ray method. Figure 5.1 shows typical spectra used for calibration and for demonstrating the effects of neutron-gamma ray separation. This spectrum was measured by SIOL as part of their detector-calibration procedure. The sum spectrum was acquired in a mixed neutron-gamma-ray field using a Pu-Be neutron source that emits both gammarays and neutrons. In the sum spectrum, the detector electronics is not tuned to separate neutrons and gamma-rays. The stilbene detector mainly contains C and H atoms and hence has a very small cross section for the photoelectric effect. Therefore the gamma-ray spectrum consists mainly of a sum of overlapping Compton spectra. The highest Compton edge corresponds to the Compton of the E = MeV line from the 12 C (n, n ) 12 C reaction and is situated at about 4.2 MeV. Figure 5.1 also shows the detector response to 152 Eu, which emits gamma-rays predominantly at, E = 0.122, 0.344, 0.779, 0.964, 1.086, and MeV. The highest Compton edge is caused by the E = MeV line and is situated at E = MeV. The neutron spectra shown in figure 5.1 were acquired with a Pu-Be source but in this case the neutron-gamma ray separation was switched on at SIOL Pu-Be n+ 152 Eu source Pu-Be n-only E Figure 5.1: Typical spectra used for energy calibration and for demonstrating the effects of neutron-gamma ray separation (supplied by SIOL, Kharkov, Ukraine). In the design of the NuPulse instrument, the fast-neutron spectra were originally only intended to monitor the flux of the D-T neutron source. As will be shown in this chapter, it was not expected to use the instrument for extracting element information from the analysis of fast-neutron spectra. As a consequence only the initial energy calibration was available to us. In the field, the calibration parameters may not be the same as in the laboratory because of various environmental measuring 104

4 conditions. Unfortunately we had no other choice than to use calibration parameters provided by SIOL. 5.3 Experiments in test facilities NuPulse test facilities at the KVI are described in section 3.9. Unlike for gamma-ray tests, neutron tests can only be conducted if the neutron generator is switched on. The fast-neutron system was tested in both the concrete water tank and the KVI borehole facility. For the water tank measurements, two conditions were used: an empty tank and a water-filled tank with the corresponding spectra presented in figure 5.2. The spectra are divided into 3 regions. A clear distinction is observed for an empty tank and for a water-filled tank for neutron energies E n < 12 MeV. For that reason area 1 and 2 were chosen; area 1 yields more or less the same neutron spectra for both conditions whereas area 2 shows marked differences. Area 2 is the region with partly moderated neutrons, hence in this area the result shows a higher count rate for the water-filled tank compared to an empty tank. Interactions in an empty tank are only limited to nuclei that make up the concrete walls, soil beneath the tank and the tool (more specifically the detector materials). Although the detector material also contains hydrogen, the flux of neutrons scattered back into the detector is considerably lower compared to the water-filled conditions. Concrete materials are mostly composed of heavy nuclei compared to hydrogen, hence neutrons will transfer a small fraction of their energy to the backscattered recoil nuclei and continue their way, sometimes with no further interaction that contributes to the detector signal. In area 1, the spectra mainly result from neutrons that have undergone hardly energy loss in their collisions before reaching the detector. Given the conditions that the neutron generator will emit neutrons almost isotropically, some neutrons will reach the detector directly from the source without too much energy loss. Neutrons reaching the detector in an indirect way would have lost energy by scattering. Hence the events in area 1 are predominantly the direct neutrons from the source and this component is to a large extent independent of the conditions around the detector. Hence the events in area 1 can be used to monitor the neutron flux produced by the generator. By monitoring the intensities in this region over time, one is able to notice if the source strength changes significantly. Area 3 is a low-energy subsection of area 2 and shows sharp peaks of which the intensity and position was found to depend on the conditions around the detector. The interpretation of results in area 3 will be given in section 5.4. In figure 5.3, the results from the borehole measurement are compared to the water-filled tank. Area 2 in figure 5.3 shows a deviation between the borehole spectrum and the water tank spectrum. The analysis of area 2 shows that the count rates in the borehole measurements are on average 40% of those in the water-filled tank. Since the water content in the borehole matrix was unknown at the time, the about 40% cannot be proven to be representative of the water content (more specifically the hydrogen content) in the borehole sediments. The value, however, is roughly consistent with expectations for water-saturated clay. Observations of area 2 are consistent with a reduced fraction of neutrons undergoing backward scattering. In the interaction with hydrogen, the presence of SiO 2 causes most neutrons to scatter more forwardly and hence escape detection. In principle the reduction factor in 105

5 FAST-NEUTRON SPECTRA ANALYSIS intensity may be used as a way to measure the hydrogen content of geological formations. However this requires a calibration set-up in which the water content can be regulated in a controlled way. In principle the water tank could have served as such a facility, but in view of the time-limited tasks in the NuPulse programme this aspect could not be further pursued. 5.4 Low-energy region, a mystery explored Figure 5.4 shows the low-energy region of the spectra in figures 5.2 and 5.3 (i.e. area 3) of the fast-neutron spectra obtained with the instrument for the empty tank, water-filled tank and the borehole at the KVI. Similar to the preceding spectra, the horizontal scale is given as an approximate energy scale, based on the SIOL calibration parameters. By the time it was realised that the low-energy region may contain compositional information, the energy calibration could not be checked or redone any more. Therefore it became a challenge to interpret this part of the spectra and consequently we were forced to derive a more precise energy calibration. This is unfortunate because we otherwise could have switched off the neutron-gamma ray separation and used gamma-ray sources to calibrate the energy scale in this part of the neutron spectrum. This lack of energy calibration imposes serious boundary conditions to our analysis, but was unfortunately an inherent attribute of the EU collaboration in the NuPulse project. Initially we focus on these three spectra and investigate if they can be interpreted in a consistent way. We start from the assumption that the three spectra exhibit the same neutron peaks, but with their individual intensities due to their difference in environment. The continuum part is assumed to be the tail of the neutron slow-down reactions as described before (area 2 in figures 5.2 and 5.3). The intensity of the continuum hence reflects mainly the amount of water in the environment. The spectra in figure 5.4 clearly exhibit this feature. All three spectra show a cut-off at about 0.9 MeV. This cut-off is likely due to a discriminator setting in the neutrongamma ray separator that was fixed during manufacturing at SIOL, Kharkov. In the following we consider two options as a way to understand our observations in the low-energy region as shown in figure 5.4. The initial assumption is that the peaks as observed are due to alpha particles produced by interactions of neutrons with the detector crystal materials (C 14 H 12 ). The well-defined peaks shown in figure 5.4 are less likely to originate from recoil protons formed by neutrons interacting with hydrogen nuclei in the crystal. In that case one would expect no sharp peaks but mainly neutron-hydrogen interactions, and a typical neutron spectrum that shows similar features as presented in figure 2.6 of section 2.6 (Krane, 1988). Alpha particles are relatively massive, hence they are unlikely to travel long distances in materials. Larmash (2001) reported that a 5-MeV alpha particle has a range of about cm in aluminium. If the alpha-particles are present, the alpha particles that deposited their energies in the stilbene detector should have been created very close to the surface of the active crystal or inside the crystal. The two possible reactions for the production of alpha particles in the detector crystal are the 12 C(n, ) 9 Be and 12 C(n,n )3 reactions. Both reactions have a significant cross section for neutron energies above 9 MeV (Knoll, 2000). The maximum -energy for the reactions 12 C(n, ) 9 Be and 12 C(n,n )3 is about 9 and 7 MeV, respectively. Applying 106

6 Count rate (cps) Count rate (cps) the quenching factor of about 6 for alphas, the expected energy deposited for the peaks will be around E = 1.42 and 1.15 MeV, respectively Empty tank Water-filled tank E n Figure 5.2: Fast-neutron spectra acquired for the KVI concrete tank facility with an empty and a water filled tank. An approximate energy scale (based on the SIOL parameters) is used for the horizontal axis Water-filled tank Borehole E n Figure 5.3: Fast-neutron spectra measured in the KVI tank filled with water and in the KVI borehole. An approximate energy scale (based on the SIOL parameters) is used for the horizontal axis. 107

7 Count rate (cps) FAST-NEUTRON SPECTRA ANALYSIS Empty tank Water-filled-tank Borehole E n Figure 5.4: Low-energy region (region labelled 3) of spectra presented in figures 5.2 and 5.3. An approximate energy scale is used for the horizontal axis. Despite the lack of precise energy calibration of the detector in the three spectra in figure 5.4, the rough energy scale presented on the horizontal axis is well within the tolerable uncertainty. All three spectra show a variation in peak intensities and shapes. The spectrum for the empty tank represents the true interactions of neutrons with the detector materials (more specifically the carbon), but only with a direct neutron flux of mainly mono-energetic neutrons. Scatterings of neutrons in the water-filled tank and the borehole will increase the neutron flux reaching the detector with a broad range of neutron energies. Thus, in comparison with a water-filled tank or a borehole to the empty tank spectrum the 12 C(n,n )3 reaction is more favourable than the 12 C(n, ) 9 Be reaction. Consequently, the assumed 1.15 MeV peak will show relatively higher intensities than the 1.42 MeV peak in the water-filled tank and borehole spectra, compared to the empty-tank spectrum. From figure 5.4, the highpeak and the low-peak positions demonstrates that the assumption is consistent with the data. In the above investigation, with proper energy calibration of the detector, the results could be used also to correctly verify the energy calibration of the detector and by intensities to confirm the moderating power of the formation. As a second possible interpretation of these low-energy neutron peaks we propose a reaction mechanism in which neutron-unstable states are populated which subsequently decay by emission of neutrons with a well-defined energy to stable states in a final nucleus. Since these neutrons have low energies, they transfer all their energies to the proton in the crystal. In all of our test environments, the nuclei H and C (stilbene (C 12 H 14 ) detector), O and Si (H 2 O and SiO 2 ) are dominantly present in various quantities and ratios. The three spectra measured in our setup are used to test the above assumptions together with the other two spectra acquired with the same detector system in Finland by the Finnish Geological Survey (GTK) for a water-filled drill hole in rock and for 108

8 urea ((NH 2 ) 2 CO) bags packed around the detector and neutron-generator components. Photographs in figure 5.5 show the experimental settings in Finland. Figure 5.5(a) shows the urea setting. To minimise the neutron scattering from the ground for the urea set-up, the experimental setting was placed at a height of about 1.5 m with wooden support as shown in the photograph. For the drill-hole experiments, the tool was lowered to a depth of about 300 m from the top surface as shown in figure 5.5(b). In these five experiments, neutron reactions with 14 N, 12 C and 16 O are expected to be dominant whereas preliminary tests showed no contribution from the 28 Si nucleus. (a) (b) Figure 5.5: Photographs of the experimental set-ups in Finland: (a) the urea experiment and (b) the drill-hole settings. Table 5.1 list the possible neutron reactions of the nuclei (column 1) of interest together with the Q-values (column 2) for the given reactions. Since the maximum neutron energy of the D-T reaction is about 14.0 MeV, only reactions with Q-values > -14 MeV have been considered. The decay reactions in column 3 result from the preceding reactions in column 1 and their Q-values are shown in column 4. For every reaction chain in each row, also the sum of the Q-values (column 5) should be larger than -14 MeV to be considered likely to occur. Also included in table 5.1 are the Coulomb barriers (E c ) that the charged ejectiles would have to overcome. Except for the 14 N(n,n ) 14 N reaction, all other possible reactions also have charged ejectiles prior to neutron emission. Hence the Coulomb barrier will also influence the reaction probabilities. Table 5.1: Possible neutron reactions with prevailing light nuclei. Reactions Q-values Reactions Q-values Total Q- values E c 12 C(n, ) 9 Be Be 8 Be + n N(n, ) 11 B B 10 B + n N(n,p) 14 C C 13 C + n N(n,n ) 14 N 0 14 N 13 N + n O(n, ) 13 C C 12 C + n O(n,p) 16 N N 15 N + n Si(n, ) 25 Mg Mg 24 Mg + n

9 FAST-NEUTRON SPECTRA ANALYSIS 12 C(n, n) 8 Be reaction The maximum -particle energy E ~ 8.5 MeV. The maximum -particle max energy to excite a neutron-unstable states which yields E n ~ 1 MeV is about 6.5 MeV. The Coulomb barrier for the -particles is about 3 MeV, so we expect that this reaction may take place. 14 N(n, n) 10 B reaction The maximum -particle energy E ~ 14.0 MeV. The maximum -particle max energy to excite a neutron-unstable states which yields E n ~ 1 MeV is about 1.5 MeV. The Coulomb barrier for the -particles is about 4 MeV, so we expect that this reaction will not have much strength. 14 N(n,pn) 13 C reaction The maximum proton energy E p ~ 14.8 MeV. The maximum proton energy max to reach a neutron-unstable state which yields E n ~ 1 MeV is about 6.5 MeV. The Coulomb barrier for the protons is about 3 MeV. In view of the large difference between proton decay to the ground state and the neutron-unstable states, it is questionable if this reaction will be observed. The situation is similar to the 14 N(n,2n) 13 N reaction. 14 N(n,2n) 13 N reaction For the 14 N(n,2n) 13 N reaction, the Q-value = MeV limits the reaction to the excitation of the low-energy states in 14 N. Therefore it is uncertain if the 14 N(n,2n) 13 N reaction will have enough probability to be observed. 16 O(n, n) 12 C reaction The maximum -particle energy E ~ 12.0 MeV. The maximum -particle max energy to reach a neutron-unstable states which yields E n ~ 1 MeV is about 6 MeV. The Coulomb barrier for the -particles is about 5 MeV. So in view of the large difference between -particle decay to the ground state and the neutron-unstable states, this reaction is is similar to the 14 N(n,2n) 13 N and 14 N(n,pn) 13 C reactions and it is unlikely that this reaction will be observed. 16 O(n,pn) 15 N reaction The maximum proton energy E p ~ 4.6 MeV. The maximum proton energy max to reach a neutron-unstable states which yields E n ~ 1 MeV is about 0.5 MeV. The Coulomb barrier for the protons is about 3 MeV. At first sight the reaction seems unlikely but since the difference in proton energies between the ground state and the neutron-unstable state is relatively small, the observation of this reaction is questionable and will depend on the structure of the states involved. 28 Si(n, n) 24 Mg reaction The maximum -particle energy E ~ 11.5 MeV. The maximum -particle max energy to reach a neutron-unstable states which yields E n ~ 1 MeV is about 4 MeV. The Coulomb barrier for the -particles is about 6 MeV. In view of the large 110

10 difference between the ground state transition and transitions to neutron-unstable states, this reaction is very unlikely to occur. Table 5.2 lists the various sample materials. For the borehole and drill-hole environments, the sample is mainly characterized by water since the silicon does not contribute. The samples column is followed by the neutron-decaying states specified by nucleus and excitation energy. The resulting neutron energies are listed in the third column. The last column shows the FWHM for the given neutron-decaying states. With the probable reactions shown in table 5.1 and the neutron energies in table 5.2, an attempt was made to derive a new energy calibration. Using MCNPX incorporated cross section data libraries, we simulated a stilbene block to determine the average mean free path of a 1-MeV neutron and we found a value of about 2.6 cm. The NuPulse stilbene scintillator crystal used for the experiment had both length and diameter of about 5.5 cm. The size of the crystal was large enough for the neutrons to interact and transfer all or most of their energies to the protons. The light output of these reactions is such that all the energy of recoil protons is received without quenching. Figures 5.6 to 5.10 present the measured spectra indicating the peak position with the proposed new energy estimate. The horizontal axis of the spectra presented in figures 5.6 to 5.10 is given in channel numbers. The peak matching was based on a best possible linear fit for the channel number and the expected neutron energy. The data points in relations between figures 5.6 to 5.10 are using the B-spline function to guide the eye. In the figures (5.6 to 5.10), the new calibration settings show that each environment can be explained in a consistent way. By examining the low-energy region, one is able to match the results to the environment in which the experiment was conducted. Further optimisation of the system to check the validity of the proposed energy calibration will improve the understanding and interpretation of the low-energy neutron spectra. Table 5.2: Proposed neutron energies for the possible reactions in table 5.1 as matched to the experimental data presented in figures 5.6 to Sample Urea Empty tank Water Neutron-decaying state E n FWHM 16 N (E x =3.35 MeV) Be (E x =2.78 MeV) Be (E x =3.05 MeV) C (E x =9.80 MeV) Be (E x =2.78 MeV) Be (E x =3.05 MeV) N (E x =3.52 MeV) Be (E x =3.05 MeV)

11 Count rate (cps) Count rate (cps) FAST-NEUTRON SPECTRA ANALYSIS Be (0.99 MeV) Be (1.23 MeV) 14 C (1.51 MeV) 0.4 Urea N (0.81 MeV) Figure 5.6: Proposed interpretation of the low-energy neutron spectrum for the urea experiment in Finland. Note the horizontal scale is presented in channel numbers N (0.97 MeV) GTK-Drill hole (300m deep) Be (1.23 MeV) Figure 5.7: Proposed interpretation of the low-energy neutron spectrum for the GTK drill-hole experiment in Finland Note the horizontal scale is presented in channel numbers. 112

12 Count rate (cps) Count rate (cps) N (0.97 MeV) KVI-Borehole (40m deep) Be (1.23 MeV) Figure 5.8: Proposed interpretation of the low-energy neutron spectrum for the borehole experiment at the KVI. Note the horizontal scale is presented in channel numbers N (0.97 MeV) 15 Water-filled tank 10 9 Be (1.23 MeV) Figure 5.9: Proposed interpretation of the low-energy neutron spectrum for the water-filled experiment at the KVI. Note the horizontal scale is presented in channel numbers. 113

13 Count rate (cps) FAST-NEUTRON SPECTRA ANALYSIS 5 4 Empty tank 3 9 Be (0.99 MeV) 9 Be (1.23 MeV) Figure 5.10: Proposed interpretation of the low-energy neutron spectrum for the empty experiment at the KVI. Note the horizontal scale is presented in channel numbers. In the next step, in order to test our assumptions, we will try to analyse the data sets provided by our NuPulse partner from Cyprus. They measured their spectra with another available D-T source and stilbene detector. The spectra from Cyprus were acquired using the FM1 detector system whereas for the KVI and Finland data the FM3 was used; for a more detailed description of the detector modules see also section 3.7 (table 3.1). We focused on the two spectra that were acquired in the urea and hydrocarbon (polyethylene; C 2 H 4 ) environments. The experimental settings for the hydrocarbon experiment in Cyprus are shown in figure The intention was to verify if the results are consistent with the plausible explanation presented for figures 5.6 to The polyethylene spectrum in figure 5.11 was considered to resemble the empty-tank spectrum of the KVI, since in the empty-tank experiment mainly reactions with stilbene (C and H) are expected. The polyethylene spectrum is consistent in the peak shape and even the peak intensities with the empty tank spectrum. It should be pointed out that the good correspondence between the polyethylene experiment at Cyprus and the empty water tank experiment at KVI may also indicate that the amount of hydrocarbon material used at Cyprus may have been too small. Figure 5.12 shows both the urea spectrum taken at Cyprus and, inserted the spectrum from Finland (i.e. figure 5.6). Several attempts were made to identify the peaks of figure 5.12 and explain the reaction processes as it was done for figure 5.6. We concluded that the peaks in figure 5.12 do not seem to be reliable for further fitting and analysis. We believe that the urea spectrum from Cyprus should have more or less the same features as the one taken in Finland regardless of using other detector systems in the two experimental settings. In conclusion of the hydrocarbon experiment, we assume that the discrepancy shows that the amount of material 114

14 Count rate (cps) Count rate (cps) surrounding the detector may have been too small. For the Finland urea experiment (see figure 5.5), the amount of urea used was large as compared to the Cyprus experiment, (see figure 5.13). The schematics in figure 5.13 show that about 3cm thick material (in this case, urea) was used at Cyprus whereas in Finland an estimated layer thickness of about 10 cm was used. The urea spectrum (figure 5.12) is not well understood. The experiment was carried out in the same period and under similar conditions as the C 2 H 4 experiment and both are compared in figure The low-energy cut-off for both spectra is consistent, hence any spectral shift that might have occurred becomes unlikely. Given the results of figure 5.12, it is clear that the urea spectrum seems not to be reliable. If the spectrum is caused by urea, the full-energy peaks are at about half the expected ones. Moreover, if the contribution of the Cyprus urea would have yielded a spectrum at lower energies, the spectrum should still have shown the peaks belonging to the reaction in the detector. Unfortunately the experiment could not be repeated. More experimental data containing heavy element materials was provided by our Greek partner. In figure 5.15, experimental results for metallic zinc, aluminium and iron are shown. The detector system (FM1) used for the Cyprus measurement was deployed also in the Greek experiments. For the three metals spectra were obtained with peaks at different channel numbers and with element-depending intensities. A pictorial view of the experimental settings in Greece is shown in figure 5.16 for the iron and zinc measurements. The spectra for Zn as well as Fe (in figure 5.15) are distinctly different from the spectra obtained in the water tank, boreholes and urea at the KVI and GTK, Finland. The shape of the Al-spectrum is similar to that obtained with an empty tank but a closer look reveals that the distance between the peaks in Al is larger and also the count rate is higher. The cut-off energies for the three spectra are not the same. A further analysis of the data was not pursued as these clearly show that a proper control of the experimental settings is required to be able to interpret the data in a quantitative method MeV Empty MeV 0.99 MeV MeV (C 2 H 4 ) Figure 5.11: Low-energy spectrum obtained in Cyprus for hydrocarbon materials. The insert shows the spectrum (fig. 5.10) for the empty tank at the KVI. 115

15 Count rate (cps) Count rate (cps) FAST-NEUTRON SPECTRA ANALYSIS Urea-F MeV 1.51 MeV MeV MeV Urea-C Figure 5.12: Low-energy spectrum from Cyprus for urea material together with the urea spectrum (insert; fig. 5.6) from Finland. Figure 5.13: A pictorial view of the Cyprus experimental settings for the urea and hydrocarbon experiments. In the top panel, the space occupied by the detectorneutron source is shown and the material is filled to the outer cylinder and the dimensions illustrated in the bottom panel. 116

16 Count rate (cps) Count rate (cps) MeV Urea C 2 H MeV Figure 5.14: The urea and hydrocarbon spectra from Cyprus measurements Zn Al Fe Figure 5.15: Low-energy spectrum obtained in Greece for zinc, aluminium and iron. The vertical dashed lines indicate the cut-off for Fe and Zn at channel 39 and Al at channel

17 FAST-NEUTRON SPECTRA ANALYSIS NuPulse probe NuPulse probe Figure 5.16: A pictorial representation of the experimental settings for the iron (top) and zinc (bottom) measurement in Greece. In the photographs the NuPulse probe is shown. 5.5 Conclusions The present results clearly show that in the middle energy region (area 2 of figures 5.2 and 5.3), the intensity seems to depend on the hydrogen content for a given environment. To develop this part of the spectrum into an analytical tool requires a stepwise approach with a fixed set-up and a set of materials ranging in hydrogen content. The low-energy neutron peaks present a novel phenomenon (International Patent, 2007). Despite the absence of a reliable energy calibration, we have shown that part of the data can be described in a consistent way in terms of delayed neutronemission. The neutron spectra are clearly depending on the environment and surrounding materials. Despite the major shortcomings in the experiments, the data indicate the potential of this low-energy neutron region could be developed into a tool to determine elemental composition. The data show that this spectroscopy does not only hold for light elements, but also for metals such as zinc and iron. One of the exciting features comes from the urea spectra. In developing instruments for drugs, landmines and explosive detections, it has been extremely difficult to detect nitrogen, the distinctive element in explosives and drugs. The urea 118

18 data clearly indicate that the low-energy neutron spectroscopy may offer new opportunities. A dedicated research activity will be needed to produce a workable instrument. The electronics of the present system has to be redesigned such that energy calibrations can be made in the field. A thorough research and development scheme has to be set-up and executed to validate and calibrate the instrument under laboratory conditions and in field experiments. Thus far we have not been successful in simulating the low-energy neutron spectra. Programmes like MCNPX require further optimisation to simulate these types of experiments. 119

19 FAST-NEUTRON SPECTRA ANALYSIS 120