EPITHERMAL AND FAST NEUTRON RADIOGRAPHY USING PHOTOLUMINESCENT IMAGING PLATES AND RESONANCE AND THRESHOLD ACTIVATION DETECTORS 1

Transcription

1 The 12 th International Conference of the Slovenian Society for Non-Destructive Testing»Application of Contemporary Non-Destructive Testing in Engineering«September 4-6, 2013, Portorož, Slovenia More info about this article: EPITHERMAL AND FAST NEUTRON RADIOGRAPHY USING PHOTOLUMINESCENT IMAGING PLATES AND RESONANCE AND THRESHOLD ACTIVATION DETECTORS 1 J.J. Rant 1, M. Balasko 2 1 Ret. from J. Stefan Institute, Ljubljana, Slovenia for correspondence: joze.rant@ijs.si 2 KFKI-AEKI, Budapest Neutron Center, Budapest, Hungary ABSTRACT Epithermal and Fast Neutron Radiography (FNR) are complementary to the conventional radiography with high energy gamma-rays or brems-strahlung radiation and which iarecommonly used for the inspection of thick metal objects. The advantage of using high energy neutrons for radiographic examination of thick metal objects is their higher penetration power and improved contrast sensitivity. Rather low detection sensitivity of direct and indirect epithermal fast neutron imaging methods using photographic films often poses a problem for FNR for NDT of thick objects. Resonance and fast neutron imaging method based on photoluminescent imaging plates (IP) in combination with activation detectors as converter screens in transfer exposure technique was successfully tested at the Budapest KFKI research reactor. The threshold activation detectors were the reactions 115 In(n, n, ) 115m In, 64 Zn(n,p) 64 Cu, 56 Fe(n,p) 56 Mn, 24 Mg(n,p) 24 Na and 27 Al(n.α) 24 Na. These threshold reactions cover the fast neutron energy region between 0.7 MeV and 12 MeV. The detectors for resonance neutron detection were activation reactions 115 In(n,γ) 116m In, 3 Cu(n,γ) 64 Cu and 164 Dy(n,γ) 165 Dy. The very high sensitivity of IP, the linearity of their response over 5 decades of exposure dose and the high dynamic digitalisation latitude (12 bit) enable practical EN/FN radiography using neutron beam of a research reactor (Φ n 10 8 n/cm 2 s, R Cd 2). The image quality of novel FNR method is comparable to the quality one encounters in thermal NR. 1. Introduction Radiography with thermal neutrons in the energy range of 0.01 ev to 0.3 ev is a well established non-destructive method [1]. Neutrons of epithermal energies (0.3 ev to 10 kev) and fast neutrons (10 kev to 20 MeV) are used for neutron radiographic examinations in much smaller extent. However, the use of fast and epithermal neutrons for neutron radiography is becoming increasingly attractive. The reasons for the increasing interest for fast neutron radiography (FNR) and epithermal neutron radiography (ENR) are: 1 This research was performed with the support of the COST 524 project "Neutron Imaging for the Detection of Defects in Materials". This support and the support of the KFKI Budapest Neutron Centre is gratefully acknowledged. 251

2 Neutrons of higher energies, in particular 14 MeV neutrons, can penetrate thick layers of bulk materials (more than 10 cm), offering the possibility to inspect larger objects [2-5]. The radiographic contrasts obtained by FNR are similar and comparable to those obtained with high energy X- and gamma-ray radiography and in general are smaller than with thermal neutron radiography. Poorer discrimination is offset by broader contrast latitude and low atomic number materials can simultaneously be observed with heavier metals [4]. However, the neutrons at the lower portion of fast neutron region (< 1 MeV) can offer radiographic contrast different from either X-ray or thermal neutron radiography [2]. Availability of mobile or even portable radioactive and modern accelerator based fast neutron sources [6-8] in addition to many research reactor based NR facilities, enables industrial applications of FNR and also ENR. The increased interest for FNR is promoting also the search for more efficient fast neutron image detectors. For the detection of fast neutron images direct techniques based either on SSNTDs [9] or scintillating screens as well as transfer techniques based on activation detectors and radiographic films were in use [1-3]. Rather low detection sensitivity of the film as well as of SSNTD based fast neutron imaging methods and also rather poor inherent image contrast of SSNTD posed a problem for the fast neutron radiography in the for the NDT interesting energy region 1-15 MeV. Direct techniques of fast neutrons using suitable p recoil converters in combination with IP or scintillating screens are known [10-13]. For more efficient detection of fast neutrons the use of novel highly sensitive photo-luminescent imaging plates (IP) in combination with threshold activation detectors as converter screens in transfer imaging technique were proposed and used by the authors [14-19]. 2. Inherent properties of transfer activation detectors in ENR/FNR The FNR using transfer technique with activation detectors has several useful properties In addition to its inherent insensitivity against γ-ray background: The threshold detectors are not sensitive for neutrons below the threshold energy of relevant nuclear reaction and hence are not sensitive for scattered neutrons moderated below threshold energy. The resonance detectors are sensitive in particular for the neutrons in the relevant resonance energy regions (energy windows) Each threshold reaction has his own characteristic energy window of sensitivity, inside which the neutrons contribute most to the activation of the detector. This energy window depends on energy dependent reaction cross-section and neutron spectrum. The neutron absorption and scattering effects in the investigated objects can significantly distort the initial fast neutron spectrum and the sensitivity regions depend on the material composition and dimensions of the object. In a single neutron exposure a set of foils of different threshold or resonance activation detectors can be irradiated providing simultaneous fast neutron radiographs through different energy windows. Simultaneously the thermal and epithermal neutrons, which are present in the neutron beam in particular when the fast neutron beams of a research reactor are used, are inducing thermal and resonance neutron absorption reactions in the converter material. This can be sometimes of an advantage, however often the unwanted 'parasitic' activities have to be suppresses by filtering out the thermal neutrons using Cd shields and/or by selecting appropriate time of neutron irradiation, waiting time before exposing the image detector and the exposure time. The reaction cross-sections of threshold detectors are in a millibarn region and orders of magnitude smaller then the activation cross-sections of conventional converter materials for 252

3 thermal neutron radiography. A problem encountered in FNR using radiographic films as image recorders is an adequate fast neutron beam intensity to produce quality fast neutron image. According to Berger for FNR using 14.5 MeV neutrons in a single screen transfer technique with common threshold detectors (reactions in Cu and Al) and radiographic films the required fast neutron flux to achieve moderate film density 1.5 at still practical exposure conditions (30 min irradiation time, waiting time 2 min and 'overnight' film exposure) should be at least of about n/cm 2 s [3]. The use of novel IPs instead of radiographic films offers several benefits. Compared to radiographic films the IPs have advantages as high sensitivity of times that of radiographic films, high dynamic range in exposure dose (more than 6 orders of magnitude), high intrinsic spatial resolution, large area and negligible spatial distortion, reusability and digital output [18]. 3. Object of the experiments The aim of the experiments to be reported here is to study some characteristics of ENR and FNR important for typical industrial non-destructive examination problems: 1. Study of contrast sensitivity of detecting thickness changes in a thick step wedge objects made of iron (5 cm thick, 0.5 cm steps), lucite (2 cm 1 cm, 1 mm steps and 0.5 cm/0.01 cm). The step wedges were imaged although through 2.5 cm thick steel plate. 2. To estimate effective attenuation coefficients of the Cd filtered research reactor neutron beam in iron, averaged over the sensitivity region of activation reactions contributing to the induced activity of converter screens made of In, Dy, Cu, Fe, Zn and Al of natural composition. 3. Examination of thick typical industrial objects for demonstration purpose First results of preliminary experiments (10 effective days of experiments at KFKI research reactor) are reported here. 4. Experimental Fast neutron source: Neutron radiography facility of the KFKI 10 MW research reactor in Budapest [20] Φ th = n/cm 2 sec R Cd ~ 1.8, collimated n beam L/D = 170 Neutron beam filtered by 0.5 mm Cd plate. Neutron converter screens: In (15 X 15 cm, 0.01 cm and cm), purity 99.99%, converter screen for TNR, Dy (12 X 12 cm, 0.01 cm), converter screen for NR and ENR, purity %, Cu (20 x 20 cm, cm), technical metal, (5 x 10 cm, 0.02 cm), threshold activation detector, purity 99.99%, Fe (20 x 20 cm, o.02 cm), technical metal, Zn ( 5 X 10 cm, cm), threshold activation detector, purity 99.99% Al (20 x 20 cm, 0.02 cm), technical metal, (10 X 10 cm, 0.01 cm), threshold activation detector, purity 99.99% In addition the converter foils were shielded by Cd (0.5mm) cover. Objects: Thick Fe step wedge, 5cm thick, 10 steps, 5 mm step Thin Fe step wedge, 5mm thick, 10 steps, 0.5mm step Thin lucite step wedge, 5mm thick, 10 steps, 0.5mm step 253

4 2.5cm thick Fe plate as additional shield Industrial objects: Pneumatic consumption meter, lower part of stainless steel 4cm thick, upper part covered by Al casing and containing debris and corrosion products mixed with oil remnants Small electric motor ENRWG BPI (beam purity indicator) standard Thick ( 15cm) visco car clutch, partially filled with oil/ Imaging plate detection system: The imaging system used consists of imaging plates, FUJI BAS 1500 laser scanner and Raytest UV IP eraser. As imaging detectors 5 different types of FUJI photoluminescent imaging plates were used ranging from highly sensitive Gd doped IP-ND, IP type SR, type UR, BAS-III IP to rather insensitive IP-TR intended for H 3 autoradiography. The BAS 1500 He-Ne laser (632.8 nm) scanner with a pixel size 0.1 mm X 0.1 mm, scanning speed 14 µs/pixel and 12 bit digitalisation (4096 grey levels) was used.. The inherent resolution of the system is lower than the pixel size, only about 0.2 mm. The digital images are handled, processed and analysed by a personal computer and printed by Epson Stylus Photo printer. The image handling, processing and analysing software is TINA 2.0 supplied by Raytest (Germany) and Photoshop ADOBE 3.0. Experimental procedures in ENR/FNR Few trial experiments in FNR at the NR facility in the radial beam channel No.1 of the Budapest KFKI reactor were performed. The neutron beam was filtered by Cd plate at the outlet of the beam. The converter screens of Al, Zn, Fe, Cu and In were usually packed together in pairs and covered by Cd cover from both sides. Irradiations were performed 'overnight' and irradiation times were typically hours. Waiting period before retrieval of the irradiated screens was about 1 hour to get rid of the short-living unwanted activities in the detector screens. In case of the 115 In(n,n ) 115m In FNR detector the waiting period was greater than 15 hours in order to suppress the 54 min activity of 116m In to the level much lower than the activity of the 115m In. Therefore two exposures were made. The first exposure after few hours of waiting (cooling!) time and of about few minutes in duration produced neutron image due to the resonance neutrons at 1.46 ev and second exposure after hours with IP exposure overnight produced fast neutron image. In the case if the neutron beam has a strong component of epithermal neutrons the choice of In as fast neutron converter for FNR might not be the best solution. Since at KFKI there was no IP scanner available, the exposed IP were transferred to IJS and read out with a delay of 24 hours! 5. Results and discussion Evaluation of the contrast sensitivity 1. With ENR using In and Dy screens it was possible to detect 5 mm thickness change of least at 40 mm to 45 mm thick Fe step, Fig. 1b and 1c respectively. Even through an additional Fe plate (2.5 cm) using In converter the step at 35mm can be detected, Fig 1a. With TNR using IP-ND the 5 mm step can observed at somewhat smaller step thickness of about 35 mm. With FNR using Fe threshold detector one can observe the 5mm step at 30 to 35mm thickness, Fig. 1d. Cu screens seem to be very promising in ENR since 9 to 10 steps of Fe wedge were revealed. 254

5 Fig. 1a ENR of Fe step wedge (0.5 cm/ 5 cm) through a 2.5 cm thick Fe plate. Lucite wedge (5 mm/ 0.5 mm) and small Fe wedge (5 mm/ 0.5 mm) are detected. Transfer exposure of a FUJI type TR IP with Cd (0.5 mm) covered In converter. Fig. 1b ENR of the same objects as in Fig. 1a, the 2.5 cm Fe plate was removed. Cd covered In converter and FUJI TR type IP. Fig. 1c ENR with Dy converter and FUJI type TR IP: Fe step wedge (5 cm/, 0.5 cm) and small lucite wedge ( 5 mm/ 0.5 mm).cd (0.5 mm) filtered n beam Fig. 1d FNR with Fe converter and FUJI type ND IP Fe step wedge (5 cm / 0.5 cm) and small lucite wedge ( 5 mm/0.5 mm).cd (0.5 mm) filtered n beam 2. ENR with In and Dy could detect all steps in small 5 mm lucite step wedge, possibly even through 2.5 cm thick Fe plate. Excellent capability of ENR with In to detect hydrogen containing materials is demonstrated in ENR of consumption meter, revealing oil soaked debris and corrosion products, Fig. 3a. This were clearly detected by TNR using IP-ND. 3. ENR with In, Dy and Cu and FNR with Fe could reveal BN disks and Cd wires of the BPI standard. However, Pb disks were not detected, Figs. 3b. 3c and 3d. Fig. 3a ENR using Cd covered In converter. Neutron irradiation time16.7 min, exposure time of a Fuji type UR IP only 100 sec, waiting time of 61 min! Fig. 3b ENR with Cd covered Cu converter. Objects: consumption meter, small electrical motor and BPI NR standard. Neutron irradiation time 24.1 h, 10 min waiting time, IP BAS III exposure for 24.3 h. The IP was overexposed, 6 th reading of the IP 255

6 Fig. 3c FNR with a Cd (0.5 mm) covered Fe converter. Irradiation time 15h, waiting time 10 min FUJI BAS III type IP exposure time 23.2 h! FNR with a Cd (0.5 mm) covered Al converter. Irradiation and exposure conditions of FUJI BAS III IP same as in Fig. 3c (simultaneous irradiation and exposure) 4. Rather small effective activation cross sections of threshold activation detectors Fe and Al did not allow to achieve sufficient exposure of IP s to reduce the statistical noise to the level below the level of signal change due to thickness change in the Fe step wedge at 30 mm 40 mm thickness. For applications using FNR the Zn converters or possibly Rh converters (high price!) could be more appropriate, as indicated by one of the experiments with a narrow pure Zn converter foil. Results are summarized in Table 1. Table 1: Contrast sensitivity in detecting thickness change of an Fe step wedge (5 cm, step 0.5 cm). Qualitative comparison of results obtained by selected converters and various IP s. Converter IP type Thick Fe step wedge 5 cm/0.5 cm Thick Fe step wedge 5 cm/0.5 cm cm Fe Thin lucite and Fe wedges 0.5 cm/.05 cm ENRWG BPI IP-ND Ther. NR In(Cd) IP-TR Dy(Cd) IP-TR Cu(Cd) IP BASIII Fe(Cd) IP-SR Al(Cd) IP-SR 7-8 steps (FOD 2 cm) 9 steps (FOD 3 cm) 8 steps (FOD 3 cm) (-) * (-) * (-) * 6-7steps (FOD 9 cm) (-) * 9-10? steps 10 steps in lucite wedge, Fe wedge det. all steps in (-) * lucite wedge Fe wedge det. Lucite and Fe wedge observed + 6 steps ++ 4 steps ++ wedge detected Only lucite Pb,BN disks, Cd wires detected (-) * BN disks, Cd wires observed BN disks, Cd wires indicated (-) ++ (-) * (-) ++ detected Only BN disks (-) * Not measured + Through 2.5 cm Fe plate ++ Not sufficient exposure due to low reaction cross section 256

7 Preliminary measurements of the effective attenuation cross-section of Cd filtered neutrons in an iron step wedge object 5. Attenuation curves of neutrons within the energy range of the effective response of In, Dy, Cu, Zn and Fe converters were measured for the Fe step wedge and corresponding attenuation coefficients were experimentally determined. From these measurements a rough estimation of the effective cross-sections for the attenuation of relevant neutrons in iron were obtained. Their comparison with the tabulated neutron cross sections pertinent to the effective energy interval of their response in the unfiltered neutron beam has only indicative value for the energy range of converters. Monte Carlo calculations with a code like MCNP-4 could provide more appropriate effective energy intervals of the response of epithermal and fast neutron converters in the Cd filtered beam and attenuated by different materials of different thickness. New, more accurate experiments to measure the effective neutron cross sections of different materials as obtained by the selected converter materials are needed. Some results are presented in Fig. 2 and in the Table 2. Table 2: Effective cross sections for attenuation of epithermal and fast neutrons in iron as measured by a transfer (indirect) technique of ENR/FNR using selected converters and IP s. Converter Main resonance or Reaction threshold E thr E low E up * Data*** at E up or E thr [b] <σ tot (Fe)> Measured ** [b] IP-ND Direct method Thermal neutrons ev In 1.4 ev ev Dy ~120 ev ev Cu ~210 ev ev Zn 1.0 MeV MeV Fe 4.2 MeV MeV or * Energy range of neutrons contributing to 80% of the response of relevant reaction, given for the neutron spectrum of a Materials Testing Reactor (MTR) (unfiltered neutrons!). Data given by J.H. Baard, W.L. Zijp and H.J. Nolthenius, Nuclear Data Guide for Reactor Neutron Metrology, Kluwer Academic Publ. And CEC, EUR EN, ** Effective total neutron cross-section of iron for selected main neutron activation reactions as estimated from the atenuation curve measured for the Fe step wedge object (5cm/0.5 cm step); Estimated rel. error ~ 5-8% 257

8 *** Total neutron cross-section of Fe, data at E up or E t (for threshold detectors): V. McLane, C.L Dunford, P.F. Rose, Neutron Cross Sections, Vol 2. Neutron Cross Section Curves, BNDC, Academic Press, For converters under Cd cover ++ Only indicative data from few measuremnts 1 TNR In Dy Fe 1 Fe 2 Zn Attenuation d Fe (mm) +++ Neutron irradiation time ~ 0.5 T 1/2. The value 6.8 obtained at saturated irradiation Fig. 2. Attenuation of neutrons in iron step wedge object measured by indirect exposure technique using converter screens made of selected materials irradiated in a Cd filtered neutron beam. Broad beam geometry, FOD was approximately 2-3 cm. Data for Fe converter represent measurement with short irradiation time (~1 T1/2, curve Fe1) and measurement at saturated irradiation (~ 6.3 T 1/2, curve Fe2). Examples of ENR/FNR using transfer technique and IP 6. The full value of using epithermal and fast neutron in the inspection of thick objects was demonstrated by ENR and FNR of an automatic car clutch, partially filled with oil and 15 cm in diameter. The object is composed of two parts with very different composition and different thermal neutron attenuation properties. The upper part is relatively hollow cavity with pistons encased in an Al casing. The lower part contains dented stainless steel wheels encased in steel casing and is relatively opaque for thermal neutrons. ENR/FNR neutron images of a pneumatic consumption meter are presented in Figs. 3a-d. Both the FNR image obtained by Fe converter and EPR obtained by Cu converter have broad latitude in contrast and both dense and hollow Al parts can be observed, including few mm thick Al casing. The FNR could not reveal oil soaked debris in the upper left corner, which was clearly observed in TNR (not presented here). The TNR was of better image quality and with better image discernment. The overall exposure of FNR is not optimal and could be still improved. New examination of industrial objects, possibly containing various types of materials of importance for the industry should be performed. The objects should be selected from those, which pose a problem for conventional gamma-ray or BS radiography or other conventional NDT methods. 7. The full value of using epithermal and fast neutron in the inspection of thick objects was demonstrated by ENR and FNR of an automatic car clutch, partially filled with oil and 15 cm in diameter. Only images obtained by Cu and Fe converter in combination with a BAS III IP were capable to reveal details in the thickest part and to observe the oil level, Fig

9 Fig. 4 FNR of "visco" automatic car clutch ( 15 cm thick) and made of steel, partially filled with oil. Oil.level can be observed at the bottom (5 cm) of the clutch.kfki Budapest N Transfer exposure of a FUJI BAS III type IP with a Fe converter (0.3 mm), irradiated under Cd cover (0.5 mm) for 16.5 h. Handling time 18 min and IP exposure time 24.3 h saturated exposure conditions! 6. Conclusions 1. ENR using In, Dy or Cu converters and commercially available IP s are suitable for the examination of about 4-5 cm, possibly even 6 cm thick iron objects. 2. Thick industrial objects with complex heterogeneous structure can be relatively easily inspected with fission like neutron spectrum employing ENR/FNR with Cu or. Fe converters in the combination with sensitive IP s. 3. Preliminary measurements of the effective attenuation cross-section of Cd filtered neutrons in an iron step wedge object contributing to the response of the selected converters in ENR/FNR have been performed. 4. New experiments and Monte Carlo modeling studies of the attenuation and detection process has to be performed in order to properly determine the effective range of the converter response. 5. References 1. Berger, H., Neutron Radiography, Elsevier, Amsterdam, Berger, H., Radiography with fast neutrons, Proc. 6 th ICNT, Paper M-7, Berger, H., Image detection methods for 14.5 MeV neutrons: techniques and applications, Int. J. Appl. Rad. Isot.,21(1970) Beynon, T.D., Constantine, G., Deep penetration radiography using 24.5 kev neutrons, Nucl. Instr. Meth. Phys., A264 (1988) Richardson, A., E., Improved images in 14.5 MeV neutron radiography, Mater. Eval., April 1977, Hawkesworth, M.R., Atomic Energy Review, 15,2 (1977) Dance, W.E., Cluzeau, S., Mast, H-U., Integration of an advanced sealed-tube neutron generator into a mobile neutron radiology system and resulting performance, Nucl. Instr. Meth.,B56/57 (1991) Cluzeau. S., Huriet, J.R., New neutron generators for industry: the GENIE family, Ibid., Berger, H., Track-etch radiography: Alpha, proton and neutron, Nucl. Techn., 19, (Sept. 1973), Matsubayashi, M. et al., preliminary examination of the applicability of imaging plates for fast neutron radiography, Nucl. Instr. Meth., A 463(2001),

10 11. Mikerov, V.I. et al., Prospects for efficient detectors for fast neutron imaging, J. Asppl. Rad. Isotopes, 61(2004) Mikerov,V.I., Samosyuk, V., Verushkin, S., Detectors based on imaging plates for fast neutron radiography, Nucl. Instr. Meth., A542(2005) Rant, J., Stade, J., Imaging plates as detector for neutron radiography, Materialprüfung, 40, 1-2, (1998) Rant, J., Stade, J., Balasko, M., Kaling, M., New possibilities in Neutron Radiography with IP, Proc. 7 th European Conference on NDT, Copenhagen, Denmark, May 1998, ECNDT (1998), Rant, J., Balasko, M., Kristof, E., Stade, J., Fast neutron radiography using IP, Neutron radiography (6), Proc. 6 th WCNR, Osaka, Japan, May 1999, 16. Rant, J., Balasko, M., Kristof, M., Stade, J., Fast Neutron Radiography using IP, Proc. 8 th Int. Conf. Nuclear Energy in Central Europe 99, Embedded COST 524 Meeting Neutron Imaging to Detect Defects in Materials, Portorož, Slovenia, 6-9 September, 1999, J. Rant, M. Balasko, Some Characteristics of Epithermal and Fast Neutron Radiography Using Photo-luminescent Imaging Plates in Transfer Exposure Technique, 4 th Intern. Topical Meeting on Neutron Radiography State College, Penn., USA, June 4-6, Rant, J. Balasko, M., Radiography with Epithermal and Fast Neutrons, Summary of the Working Group meetings, COST- action 524 "Neutron Imaging for the Detection of Defects in Materials", Technical University Muenchen, Garching, September 13-14, PSI report, 2001, Balaskó, M. Sváb, E., "Dynamic neutron radiography instrumentation and applications in Central Europe" Nucl. Instr. Methods A 377,(1996) Baard, J.H., Zijp, W.L., Nolthenius, H.J., Nuclear Data Guide for Reactor Neutron Metrology, EUR EN, Kluwer Acad. Publ., Dordrecht, JEF-PC 2.0, NEA Data Bank, OECD, Paris,