- Email: [email protected]
Performance of the Bragg-edge Transmission Imaging at a Compact Accelerator-driven Pulsed Neutron Source
Available online at www.sciencedirect.com
ScienceDirect Physics Procedia 60 (2014) 254 – 263
Union of Compact Accelerator-Driven Neutron Sources (UCANS) III & IV
Performance of the Bragg-edge transmission imaging at a compact accelerator-driven pulsed neutron source Hirotaka Satoa*, Yoshinori Shiotaa, Takashi Kamiyamaa, Masato Ohnumaa, Michihiro Furusakaa and Yoshiaki Kiyanagia,b a b
Faculty of Engineering, Hokkaido University, Kita-13 Nishi-8, Kita-ku, Sapporo, Hokkaido 060-8628, Japan Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8603, Japan
Abstract Performances of the Bragg-edge transmission imaging at a compact accelerator-driven pulsed neutron source (pulsed CANS) are presented and evaluated. This technique is expected to be a new material analysis tool that can quantitatively visualize crystalline microstructural information inside a bulk material over large area with reasonable spatial resolution non-destructively. Therefore, it is expected that such new useful instrument should be installed at not only world-leading pulsed spallation neutron sources but also popular-priced CANS. For this reason, we evaluated and discussed the performances of the Bragg-edge transmission imaging at CANS for potential users. A coupled moderator is usually used to gain higher neutron flux at CANS. In such situation, quantitative imaging of crystal lattice strain and crystalline phase is not easy due to the low wavelength resolution. However, according to Monte-Carlo simulation calculation studies, it was found that an experimental setup using a decoupled moderator connected to a supermirror guide tube can solve this problem. On the other hand, in the situation using the coupled moderator, quantitative imaging of crystallographic texture and crystallite size can be carried out, but the Rietveldtype data analysis software, RITS, is necessary to evaluate reasonably low statistics data measured at CANS. Furthermore, it was found that reasonable results can be obtained by the Bragg-edge transmission imaging with the RITS code at CANS, which are consistent with results of a high-performance neutron diffraction experiment with the Rietveld analysis at a world-leading pulsed spallation neutron source. This means the Bragg-edge transmission imaging is expected to be one of the most efficient crystallographic/metallographic analysis tools for CANS. © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). © 2014 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of UCANS Peer-review under responsibility of the Organizing Committee of UCANS III and UCANS IV
Keywords: Pulsed neutron imaging; Bragg-edge transmission; Compact accelerator-driven short-pulsed neutron source.
* Corresponding author. Tel.: +81-11-706-6679; fax: +81-11-706-6679. E-mail address: [email protected].
1875-3892 © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). Peer-review under responsibility of the Organizing Committee of UCANS III and UCANS IV doi:10.1016/j.phpro.2014.11.035
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
1. Introduction The Bragg-edge transmission imaging using a two-dimensional position-sensitive detector combined with the time-of-flight (TOF) method at a pulsed neutron source is expected to be a new material analysis tool that has different properties from SEM-EBSD, X-ray/neutron scattering and synchrotron radiation microtomography. This is because this method can non-destructively deduce crystalline microstructural information (crystal structure, crystalline phase, crystallographic texture (preferred orientation), microstructure (crystallite size) and crystal lattice strain) of a bulk material over large area (~ 10 cm) with reasonable spatial resolution (~ 100 μm) from the position-dependent Bragg-edge transmission spectra. Furthermore, this method can become a quantitative evaluation technique owing to a Bragg-edge analysis software such as RITS (Rietveld Imaging of Transmission Spectra) [1-3]. For these reasons, it was decided that the first pulsed neutron imaging instruments should be constructed at world-leading pulsed spallation neutron sources, J-PARC MLF in Japan, RAL ISIS in UK, ESS in EU and ORNL SNS in USA. Development activities of the Bragg-edge transmission imaging are typical examples of contribution methodology of a compact accelerator-driven neutron source (CANS). For example, at HUNS (Hokkaido University Neutron Source) in Japan, the Bragg-edge transmission spectroscopy/imaging has been developed [1], and exported to J-PARC MLF in Japan [2,4,5]. At the Bariloche LINAC in Argentina, the Bragg-edge transmission spectroscopy has been developed [6-8], and exported to RAL ISIS in UK [9,10]. Thus, CANS can contribute to the development of new neutron technologies, and export them to worldleading neutron experimental facilities. Here, let us remember two main roles of CANS; not only promotion of such development activity for new neutron technologies (neutron engineering), but also promotion of utilization activity of various neutron experiments (materials science and industrial application). However, actually, the latter’s examples on the Bragg-edge transmission imaging are not so many although it has been already found by the first experimental demonstration at HUNS [1] that the Bragg-edge transmission imaging is also feasible at pulsed CANS. In this paper, we evaluate performances of the Bragg-edge transmission imaging at CANS in detail by using the data of the first quantitative texture/microstructure imaging experiment performed at HUNS [1]. In the discussion, we compare them with data of a high-performance neutron diffraction experiment and the Rietveld analysis performed at a world-leading pulsed neutron experimental facility, J-PARC MLF [11]. Furthermore, we also present a future plan for higher intensity/wavelength-resolution experiment at CANS, based on Monte-Carlo simulation calculation studies. By these items, the feasibility information of the Bragg-edge transmission imaging at CANS can be provided to potential users in detail.
2. Specifications of a pulsed neutron transmission imaging experiment at HUNS In this section, we present the specifications of a pulsed neutron transmission imaging experiment at HUNS, which achieved the first quantitative texture/microstructure imaging with the RITS code [1]. This experiment was carried out under the severest conditions due to high spatial resolution (800 μm), namely, small pixel size causing low neutron counting. In other words, here, we present the specification requirements for quantitative imaging of crystalline structural information at CANS; neutron source, neutron transport and neutron TOF imaging detector. 2.1. Neutron source HUNS is a pulsed cold neutron source based on the s-band 45 MeV electron linear-accelerator installed at Hokkaido University in Japan. Fig. 1 shows the specifications and the schematic view of
255
256
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
TMRA (target-moderator-reflector assembly) of HUNS. For neutron experiments, the Hokkaido University LINAC operates 1 kW (= 33 MeV × 33 μA) of the beam power, and the pulse width of the electron beam is 3 μs with the pulse repetition of 50 Hz. The neutron generation reaction used is the photonuclear reaction caused by bremsstrahlung of electrons. The target material is Pb. The total yield of fast neutrons emitted from the target material is about 1.6×1012 n/s. The high-energy evaporation neutrons are slowing down from the MeV region to the meV region through the neutron moderator. The neutron moderator used at HUNS is usually a coupled type moderator for neutron beam extraction of higher intensity [12]. The main moderator is 18 K solid methane (CH4) of 12 cm × 12 cm × t 5 cm, and the premoderator is polyethylene of 1.5 cm thickness. The neutron pulse width (FWHM of emission time of a certain wavelength neutrons emitted from the main moderator) is about 160 μs for neutrons of the 0.4 nm wavelength. This value is not so narrow because of the high-intensity type coupled moderator. The neutron reflector around the moderator consists of graphite with the dimensions of 80 cm × 80 cm × 80 cm. This volume was decided by a neutronic design based on a Monte-Carlo simulation calculation.
Electron LINAC z 33 MeV㽢33 μA = 1 kW z 3 μs pulse width z 50 Hz pulse repetition Neutron generator z Electron bremsstrahlung z Photonuclear reaction z Pb target z 1.6㽢1012 n/s yield
Polyethylene premoderator (t 1.5 cm) 3RO\HWK\OHQH
Solid methane moderator (12 cm㽢12 cm㽢 t 5 cm)
Neutron moderator z Coupled type (for higher intensity) z 18 K solid CH4 (12 cm㽢12 cm㽢 t 5 cm) z Polyethylene premoderator (t 1.5 cm) z 160 μs pulse FWHM (4ύ neutrons) Neutron reflector z Graphite (80㽢80㽢80 cm3)
10 cm㽢10 cm beam extraction
Lead target Graphite reflector
Fig. 1. Specifications and schematic view of Hokkaido University Neutron Source (HUNS).
2.2. Neutron transport Fig. 2 shows the specifications and the schematic layout of the neutron beam-line geometry for this experiment. We put specimens and a neutron TOF imaging detector at the 6 m position apart from the main moderator surface. Therefore, the neutron wavelength band width was from 0.0 nm to 1.3 nm in case of the 50 Hz pulse repetition. All collimators setup at the beam-line were opened under the condition of the 10 cm × 10 cm square. Therefore, the umbra beam size was 10 cm × 10 cm, and so-called L/D (the collimator ratio) was 60 (± 0.5° beam angular divergence). The neutron flux at the sample/detector
257
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
Sample & Detector position, and Wavelength band width z 6 m from the moderator surface э 1.3 nm wavelength band width Collimators (B4C & Boric-acid resin) z 10 cm㽢10 cm square collimators (10 cm㽢10 cm umbra beam size) Collimator ratio (L/D) z L/D = 60 (= Neutron flight path length 600 cm / Effective moderator size 10 cm) z 㼼0.5㼻beam angular divergence Neutron flux z 8.6㽢103 n/cm2/s @ sample & detector position Wavelength resolution z 2.7% (4ύ neutrons) (= Pulse FWHM 160 μs / Neutron flight time 6067 μs)
Neutron source
Evacuated beam tube Neutron
3.0 m
B4C shield
GEM detector
0.5 m
Specimens
Total flight path length : 6 m Fig. 2. Specifications and schematic layout of the neutron beam-line geometry for this experiment.
position was 8.6×103 n/cm2/s. The neutron wavelength resolution was 2.7% for neutrons of the 0.4 nm wavelength. 2.3. Neutron TOF imaging detector We used a GEM (gas electron multiplier) detector [13] for this experiment. This is because our aim was an experiment with medium spatial resolution (~ 1 mm). The GEM detector is a type of a micro pattern gas detector (MPGD). The GEM detector used in this experiment contained two 10B coated GEM foils and one electron-amplifying GEM foil. The gas used was 70%Ar and 30%CO2 under the pressure of 0.1 MPa. The pixel size (position resolution) due to the readout electrode was 800 μm × 800 μm. The detection area was 9.6 cm × 9.6 cm square composed of 120 × 120 pixels (14,400 positions). The detection efficiency was 15% for neutrons with the wavelength of 0.4 nm. The maximum neutron counting rate depending on the DAQ system was about 106 Hz (1 MHz).
3. Experimental results and comparison with a high-performance neutron powder diffractometer In this experiment, under the experimental condition described in Sec. 2, we measured Bragg-edge transmission spectra at 14,400 positions of rolled/welded α-Fe plates of 6 mm thickness. In this section, firstly, we present and discuss features of Bragg-edge transmission spectra measured at short-pulsed CANS based on the coupled moderator. Furthermore, we present results of quantitative
258
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
texture/microstructure imaging with the Rietveld-type analysis (the RITS code), and also compare them with results of a high-performance neutron powder diffraction experiment with the Rietveld analysis at a world-leading pulsed spallation neutron source. 3.1. Measured Bragg-edge transmission spectrum and its features Fig. 3 (a) shows incident/transmitted neutron spectra, and Fig. 3 (b) shows the neutron transmission spectrum derived from these two spectra (Fig. 3 (a)), measured by one pixel (800 μm × 800 μm area) of the GEM detector. The binning time width of the TOF analysis is 50 μs. The measurement time was 3.3 hours for incident beam, and 5.0 hours for transmitted beam. The measured samples were rolled/welded α-Fe plates of 6 mm thickness. Therefore, a large Bragg-edge caused by {110} diffraction of the bodycentered-cubic (BCC) crystal structure was observed. On the other hand, the data statistics were quite low, about 4% of the statistics error, due to low counts of neutrons, less than 600 counts as shown in Fig. 3 (a). This is because of not only low neutron flux at HUNS (8.6×103 n/cm2/s) but also small pixel size (800 μm × 800 μm) and low detection efficiency (15% for cold neutrons) of the detector.
(a) Raw neutron spectra (1 pixel)
(b) Neutron transmission spectrum (1 pixel) 80%
500 400
Incident neutrons
300
(3.3 hours)
200
Transmitted neutrons
(5.0 hours)
100
70% 60%
Fe {110}
Counts of neutron
600
Neutron transmission
700
50%
40% 30%
0 0.0
0.2
0.4
0.6
0.8
Neutron wavelength / nm
1.0
1.2
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Neutron wavelength / nm
Fig. 3. Neutron spectra measured by one pixel with the TOF bin width of 50 μs. (a) Raw spectra of incident and transmitted neutrons. (b) Neutron transmission spectrum derived from the Fig. 3 (a) spectra.
Furthermore, the broad neutron pulse effect due to the coupled moderator for higher neutron flux is discussed. Fig. 4 shows the data of Bragg-edge transmission spectra for cases of both high and low wavelength resolution, respectively. The data for the case of high wavelength resolution (Δλ/λ = 0.15% for cold neutrons) were measured at J-PARC MLF BL19 “TAKUMI” [14] at the 40 m position from the decoupled-poisoned moderator of JSNS (Japan Spallation Neutron Source) at J-PARC MLF. The data for the case of low wavelength resolution (Δλ/λ = 2.7% for cold neutrons) were measured by this presented experimental setup performed at the 6 m position from the HUNS coupled moderator. The samples were α-Fe plates in both cases. In the low wavelength resolution case, broad Bragg-edges were observed. Therefore, quantitative imaging of crystal lattice strain and crystalline phase is not easy for CANS using a coupled moderator. This is because correct decision of wavelength position of a Bragg-edge (strain analysis) or separation between Bragg-edges (crystalline phase analysis) is not easy. On the other hand, quantitative imaging of texture and crystallite size is not so difficult for such experimental setup because analyzed wavelength regions are far from Bragg-edges (transmission intensities between Bragg-edges). Thus, the data achieved from these experiments indicate that quantitative texture/microstructure imaging can be carried out easily for a coupled moderator at CANS. Incidentally, a methodology to perform quantitative imaging of strain and crystalline phase at CANS is presented in Sec. 4.
259
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
(a) High wavelength resolution case
(b) Low wavelength resolution case 70%
Neutron transmission
Neutron transmission
80% 70%
60% 50% 40% 30%
60%
50%
40% 0.1
0.2 0.3 0.4 Neutron wavelength / nm
0.5
0.1
0.2 0.3 0.4 Neutron wavelength / nm
0.5
Fig. 4. Comparison of Bragg-edge neutron transmission spectra between (a) the high wavelength resolution case (0.15% for 0.4 nm wavelength neutrons) measured at J-PARC MLF BL19 “TAKUMI” and (b) the low wavelength resolution case (2.7% for 0.4 nm wavelength neutrons) measured by this presented experiment.
Energy-selective (wavelength-resolved) neutron radiography also can be performed easily by this experimental setup. Fig. 5 shows wavelength-dependent neutron radiographs measured by this experiment. Two rolled α-Fe plates were set at the upper side in samples, and neutrons were transmitted along the normal direction (ND) of the rolled plates. One welded α-Fe plate was set at the lower side in samples, and neutrons were transmitted along ND of the plate. The welded zone was located along the center-line. The other welded α-Fe plate was set at the middle position in samples, and neutrons were transmitted along the rolling direction (RD) of the plate. Up to the Bragg cut-off wavelength (~ 0.4 nm), lower transmission values are distributed around the welded zone, caused by changes of texture and crystallite size due to the welding processing. It is known that such energy-selective neutron radiography can be related with the SEM-EBSD results complementarily [15]. Thus, at CANS, even if quantitative crystalline microstructural information imaging combined with the Rietveld-type data analysis is difficult, the energy-selective neutron radiography can be performed easily, and can contribute to materials science. 3.2. Quantitative visualization of texture and microstructure with the Rietveld-type data analysis For quantitative imaging of texture and crystallite size, we analyzed all Bragg-edge transmission spectrum data for every pixel by using the RITS code. The results of the imaging were already presented in Ref. [1]. We were successfully able to obtain reasonable images on texture and crystallite size with the reasonable spatial resolution of 800 μm although we carried out this experiment under the condition of quite low neutron counting, as shown in Fig. 3. Owing to the data analysis software “RITS”, we can correctly evaluate crystalline microstructural parameters of the samples at HUNS because the profile (curve) fitting analysis is applicable to low statistics data at CANS. In other words, the RITS code is reactive for low statistics data, and is useful for experiments at CANS. 3.3. Comparison with a neutron diffraction experiment at a huge pulsed spallation neutron source Furthermore, we compared the Bragg imaging results with results obtained by a high-performance neutron powder diffractometer installed at a huge pulsed spallation neutron experimental facility. The pulsed neutron powder diffractometer was iMATERIA at J-PARC MLF BL20 [16]. The Rietveld analysis software was Z-Rietveld developed at J-PARC MLF [17]. During the diffraction experiment, the power
260
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
O = 0.18 nm
O = 0.30 nm 88
0
64
40 -5
88
0
64
40
-5
0 +5 Position x / cm
-5
O = 0.40 nm 88
0
64
40 0 +5 Position x / cm
0
64
40
-5 -5
+5
88
0
64
40
-5 -5
0 +5 Position x / cm
0 +5 Position x / cm
O = 0.60 nm
Cold neutron transmission 䠄%䠅 䠄Wavelength : 0.49 nm 䡚 0.51 nm䠅
Position y / cm
Position y / cm
+5
-5
88
O = 0.50 nm
Cold neutron transmission 䠄%䠅 䠄Wavelength : 0.39 nm 䡚 0.41 nm䠅
-5
+5
0 +5 Position x / cm
Cold neutron transmission 䠄%䠅 䠄Wavelength : 0.59 nm 䡚 0.61 nm䠅
Position y / cm
-5
+5
Low energy neutron transmission 䠄%䠅 䠄Wavelength : 0.34 nm 䡚 0.36 nm䠅
Position y / cm
+5
O = 0.35 nm
Low energy neutron transmission 䠄%䠅 䠄Wavelength : 0.29 nm 䡚 0.31 nm䠅
Position y / cm
Position y / cm
Thermal neutron transmission 䠄%䠅 䠄Wavelength : 0.17 nm 䡚 0.19 nm䠅
+5
88
0
64
40
-5 -5
0 +5 Position x / cm
Fig. 5. Wavelength-dependent neutron radiographs measured by this experiment.
of the J-PARC proton accelerator was 120 kW although the full power is 1 MW, and the neutron yield of JSNS was 104 times higher than that of HUNS. The moderator of BL20 was the decoupled-poisoned 20 K supercritical 100% para-H2 moderator. 3Qc supermirror guide tube of 14 m length was installed. The distance from the moderator to the sample was 26.5 m, and the distances from the sample to the detectors were 2.0 ~ 2.3 m. The high-resolution backscattering detector bank was used. The resolution of crystal lattice plane spacing was 0.16%. The measurement time was 30 minutes for one irradiation. One irradiation can investigate only one position in a sample. In this diffraction experiment, the same samples as the Bragg imaging experiment at HUNS were measured, which is already presented in Sec. 3.1 and 3.2. Fig. 6 shows results of the comparison between the Bragg-edge transmission imaging with the RITS code at HUNS and the diffraction with the Z-Rietveld code at J-PARC MLF BL20 “iMATERIA”. Both of the results (texture degree and crystallite size) for any samples’ position/direction were corresponded within the analysis error. Correctly, only for crystallite size, 1.55 times larger values of the diffraction analysis can correspond with the results of the Bragg-edge analysis. We guess this difference is caused by neutron wavelength resolution. Higher resolution can sharpen a diffraction peak, and can weaken the extinction effect. For the crystallite size analysis, both codes use the extinction effect caused by rediffraction of neutrons inside a crystallite. Therefore, the high-resolution diffraction experiment evaluated smaller extinction effect and finer crystallites than the low-resolution Bragg imaging experiment. Thus, the Bragg-edge transmission imaging at CANS can give the reasonable data with high precision, which is consistent with a high-performance neutron powder diffractometer installed at a world-leading pulsed spallation neutron source. In other words, the Bragg-edge transmission imaging is expected to be one of the most efficient crystallographic/metallographic analysis tools for pulsed CANS.
261
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
1.0
(b) Crystallite size
Bragg imaging Diffraction
0.8
0.6
Weld/ND Weld/RD Roll/ND Roll/RD
0.4 0.2 0.0
6.0
Crystallite size / μm
March-Dollase coefficient
(a) Degree of crystallographic anisotropy (March-Dollase coefficient)
5.0
Roll/RD
4.0
Bragg imaging Diffraction×1.55 Diffraction
Roll/ND
3.0
Weld/RD Weld/ND
2.0 1.0 0.0
0
1 2 3 4 Evaluated position/direction
5
0
1 2 3 4 Evaluated position/direction
5
Fig. 6. Results of the comparison between the Bragg-edge transmission imaging with RITS at HUNS and the diffraction with ZRietveld at J-PARC. (a) Degree of crystallographic anisotropy, indicated by the March-Dollase coefficient. (b) Crystallite size.
4. Future prospect toward quantitative imaging of strain and crystalline phase at CANS Due to the low wavelength resolution of a coupled moderator for higher low-energy neutron yield at CANS, quantitative analysis/imaging of strain and crystalline phase using the Bragg-edge transmission imaging is not easy for CANS. On the other hand, in case of a decoupled moderator use, the higher wavelength resolution can be achieved, but the neutron flux becomes weaker than the coupled moderator case. Therefore, we considered an experimental setup using a supermirror guide tube for efficient neutron beam transport, combined with the decoupled moderator. Hereafter, we discuss effects of such experimental setup at CANS by using a Monte-Carlo neutron transport simulation code, McStas [18]. Fig. 7 shows a schematic layout for the simulation. This geometry is a reasonable beam-line design at HUNS. The sample/detector position was the 6 m position apart from a decoupled hydrogen moderator. At the just center position of the whole beam-line, a 3.65Qc supermirror guide tube of 4 m length was set. The open area of the tube was 10 cm × 10 cm square. In the simulation, we evaluated wavelength resolution (sharpness of a Bragg-edge) due to the decoupled moderator, neutron flux complement due to the supermirror guide tube, and increase of beam angular divergence due to the supermirror guide tube. Fig. 8 (a) shows sharpness of a Bragg-edge which will be measured by the experimental setup as shown in Fig. 7. The result is presented together with the Δλ/λ = 2.7% data of the coupled CH4 moderator setup at HUNS (the neutron flight path length was 6 m), and the Δλ/λ = 0.35% data of the decoupled supercritical 100% para-H2 moderator setup at J-PARC MLF BL10 “NOBORU” [19] (the neutron flight path length was 14 m). According to Fig. 8 (a), at CANS, a sharp Bragg-edge can be observed owing to the decoupled moderator, and the wavelength resolution of 0.8% can be achieved. This value is not so bad for the strain/phase analysis of the Bragg-edge transmission spectroscopy. Fig. 8 (b) shows a result of the neutron flux recovery owing to the supermirror guide tube. In case of only the decoupled hydrogen moderator, the neutron flux becomes quite weak, about 10% intensity of the coupled hydrogen moderator. In case of the decoupled moderator setup combined with the supermirror guide tube, the neutron flux can increase up to 150% intensity of the coupled moderator. The better intensity complement is expected in the methane moderator case than the hydrogen moderator case. Finally, Table 1 shows change of the beam angular divergence before/after setting of the supermirror guide tube, with information of its corresponding L/D value. After the set of the supermirror guide tube, the beam angular divergence increases, but is not so bad. This is because the 1 mm spatial resolution,
262
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
reasonable value for quantitative crystalline microstructural information imaging using the Bragg-edge transmission spectroscopy, can be achieved for a several-mm thickness sample directly coupled with an imaging detector (no distance between the sample and the detector). Thus, owing to the decoupled moderator setup combined with the supermirror guide tube, quantitative imaging of not only texture and microstructure but also strain and crystalline phase will be reasonably performed at pulsed CANS. Decoupler
3.65Qc supermirror guide tube
Moderator
Sample Imaging detector
Reflector
4m 6m
Fig. 7. Schematic layout of the McStas simulation for a higher wavelength resolution experiment at HUNS/CANS.
Total cross section / a.u.
0.8
Fe {110} 0.7
0.6 0.5 0.4
Δλ/λ = 0.35% (J-PARC MLF BL10)
(b) Gain of neutron intensity
Δλ/λ (6 m from decoupled HUNS) ~ 0.8% Δλ/λ = 2.7% (HUNS)
0.3
0.38
0.40 0.42 0.44 0.46 Neutron wavelength / nm
0.48
Flux ratio to coupled HUNS
(a) Sharpness of Bragg edge
2.5 2.0
1.5
Decoupled moderator w/ guide tube
1.0
Coupled moderator
0.5
Decoupled moderator 0.0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Neutron wavelength / nm
Fig. 8. (a) Sharpness of a Bragg-edge due to the decoupled moderator setup. (b) Gain of the decoupled hydrogen moderator setup combined with/without a supermirror guide tube. The gain means the ratio to intensity of the coupled hydrogen moderator. Table 1. Comparison of beam angular divergence before/after setting of the supermirror guide tube.
Without supermirror guide tube With supermirror guide tube
Beam angular divergence ± 0.5° (L/D ~ 60) ± 2.4° (L/D ~ 12.5)
5. Conclusion We evaluated and discussed performances of the Bragg-edge transmission imaging at a compact accelerator-driven short-pulsed neutron source in order to indicate the detailed feasibility of this new useful technique at a popular-priced facility for industrial use. A coupled moderator is usually used to gain higher beam intensity. Therefore, quantitative imaging of strain and crystalline phase is not easy due to broad Bragg-edges. For this problem, it was found that a high wavelength resolution experimental setup
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
using a decoupled moderator connected to a supermirror guide tube is effective, according to the MonteCarlo simulation calculation based neutronic studies. Even if the coupled moderator is used, quantitative imaging of crystallographic texture and microstructure (crystallite size) can be achieved. The important matter to analyze low statistics data is to use the data analysis software that can perform the profile (curve) fitting toward experimental data, like the Rietveld-type data analysis software, RITS. Furthermore, such software can provide quantitative data that are consistent with results of a high-performance neutron diffractometer at a world-leading pulsed spallation neutron source. Thus, the Bragg-edge transmission imaging is expected to be one of the most efficient crystallographic analysis tools for a compact accelerator-driven short-pulsed neutron source. Acknowledgements This work was partially supported by Grant-in-Aid for Scientific Research (S) from Japan Society for the Promotion of Science (No. 23226018). References [1] H. Sato, T. Kamiyama and Y. Kiyanagi, Mater. Trans. 52 (2011) 1294-1302. [2] Y. Kiyanagi, H. Sato, T. Kamiyama and T. Shinohara, J. Phys. Conf. Ser. 340 (2012) 012010. [3] H. Sato, T. Shinohara, R. Kiyanagi, K. Aizawa, M. Ooi, M. Harada, K. Oikawa, F. Maekawa, K. Iwase, T. Kamiyama and Y. Kiyanagi, Phys. Proc. 43 (2013) 186-195. [4] Y. Kiyanagi, T. Kamiyama, H. Sato, T. Shinohara, T. Kai, K. Aizawa, M. Arai, M. Harada, K. Sakai, K. Oikawa, M. Ohi, F. Maekawa, T. Sakai, M. Matsubayashi, M. Segawa and M. Kureta, Nucl. Instrum. Methods A 651 (2011) 16-20. [5] Y. Kiyanagi, T. Shinohara, T. Kai, T. Kamiyama, H. Sato, K. Kino, K. Aizawa, M. Arai, M. Harada, K. Sakai, K. Oikawa, M. Ooi, F. Maekawa, H. Iikura, T. Sakai, M. Matsubayashi, M. Segawa and M. Kureta, Phys. Proc. 43 (2013) 92-99. [6] F. Kropff and J. R. Granada, Unpublished Report CAB-1977, Institute Balseiro, Bariloche (1977). [7] J. R. Granada, J. R. Santisteban, J. Dawidowski and R. E. Mayer, Proceedings of 9th International Topical Meeting on Nuclear Research Applications and Utilization of Accelerators (AccApp09), AP/IE-02, International Atomic Energy Agency, Vienna (2010). [8] S. Petriw, J. Dawidowski and J. Santisteban, J. Nucl. Mater. 396 (2010) 181-188. [9] J. R. Santisteban, L. Edwards, M. E. Fitzpatrick, A. Steuwer, P. J. Withers, M. R. Daymond, M. W. Johnson, N. Rhodes and E. M. Schooneveld, Nucl. Instrum. Methods A 481 (2002) 765-768. [10] W. Kockelmann, S. Y. Zhang, J. F. Kelleher, J. B. Nightingale, G. Burca and J. A. James, Phys. Proc. 43 (2013) 100-110. [11] H. Sato, T. Kamiyama, K. Iwase, T. Ishigaki and Y. Kiyanagi, Nucl. Instrum. Methods A 651 (2011) 216-220. [12] Y. Kiyanagi, N. Watanabe and H. Iwasa, Nucl. Instrum. Methods A 312 (1992) 561-570. [13] S. Uno, T. Uchida, M. Sekimoto, T. Murakami, K. Miyama, M. Shoji, E. Nakano and T. Koike, Phys. Proc. 37 (2012) 600-605. [14] S. Harjo, T. Ito, K. Aizawa, H. Arima, J. Abe, A. Moriai, T. Iwahashi and T. Kamiyama, Mater. Sci. Forum 681 (2011) 443448. [15] E. H. Lehmann, S. Peetermans, L. Josic, H. Leber and H. van Swygenhoven, Nucl. Instrum. Methods A 735 (2014) 102-109. [16] T. Ishigaki, A. Hoshikawa, M. Yonemura, T. Morishima, T. Kamiyama, R. Oishi, K. Aizawa, T. Sakuma, Y. Tomota, M. Arai, M. Hayashi, K. Ebata, Y. Takano, K. Komatsuzaki, H. Asano, Y. Takano and T. Kasao, Nucl. Instrum. Methods A 600 (2009) 189-191. [17] R. Oishi, M. Yonemura, Y. Nishimaki, S. Torii, A. Hoshikawa, T. Ishigaki, T. Morishima, K. Mori and T. Kamiyama, Nucl. Instrum. Methods A 600 (2009) 94-96. [18] K. Lefmann and K. Nielsen, Neutron News 10 (1999) 20-23. [19] F. Maekawa, K. Oikawa, M. Harada, T. Kai, S. Meigo, Y. Kasugai, M. Ooi, K. Sakai, M. Teshigawara, S. Hasegawa, Y. Ikeda and N. Watanabe, Nucl. Instrum. Methods A 600 (2009) 335-337.
263
ScienceDirect Physics Procedia 60 (2014) 254 – 263
Union of Compact Accelerator-Driven Neutron Sources (UCANS) III & IV
Performance of the Bragg-edge transmission imaging at a compact accelerator-driven pulsed neutron source Hirotaka Satoa*, Yoshinori Shiotaa, Takashi Kamiyamaa, Masato Ohnumaa, Michihiro Furusakaa and Yoshiaki Kiyanagia,b a b
Faculty of Engineering, Hokkaido University, Kita-13 Nishi-8, Kita-ku, Sapporo, Hokkaido 060-8628, Japan Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8603, Japan
Abstract Performances of the Bragg-edge transmission imaging at a compact accelerator-driven pulsed neutron source (pulsed CANS) are presented and evaluated. This technique is expected to be a new material analysis tool that can quantitatively visualize crystalline microstructural information inside a bulk material over large area with reasonable spatial resolution non-destructively. Therefore, it is expected that such new useful instrument should be installed at not only world-leading pulsed spallation neutron sources but also popular-priced CANS. For this reason, we evaluated and discussed the performances of the Bragg-edge transmission imaging at CANS for potential users. A coupled moderator is usually used to gain higher neutron flux at CANS. In such situation, quantitative imaging of crystal lattice strain and crystalline phase is not easy due to the low wavelength resolution. However, according to Monte-Carlo simulation calculation studies, it was found that an experimental setup using a decoupled moderator connected to a supermirror guide tube can solve this problem. On the other hand, in the situation using the coupled moderator, quantitative imaging of crystallographic texture and crystallite size can be carried out, but the Rietveldtype data analysis software, RITS, is necessary to evaluate reasonably low statistics data measured at CANS. Furthermore, it was found that reasonable results can be obtained by the Bragg-edge transmission imaging with the RITS code at CANS, which are consistent with results of a high-performance neutron diffraction experiment with the Rietveld analysis at a world-leading pulsed spallation neutron source. This means the Bragg-edge transmission imaging is expected to be one of the most efficient crystallographic/metallographic analysis tools for CANS. © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). © 2014 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of UCANS Peer-review under responsibility of the Organizing Committee of UCANS III and UCANS IV
Keywords: Pulsed neutron imaging; Bragg-edge transmission; Compact accelerator-driven short-pulsed neutron source.
* Corresponding author. Tel.: +81-11-706-6679; fax: +81-11-706-6679. E-mail address: [email protected].
1875-3892 © 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). Peer-review under responsibility of the Organizing Committee of UCANS III and UCANS IV doi:10.1016/j.phpro.2014.11.035
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
1. Introduction The Bragg-edge transmission imaging using a two-dimensional position-sensitive detector combined with the time-of-flight (TOF) method at a pulsed neutron source is expected to be a new material analysis tool that has different properties from SEM-EBSD, X-ray/neutron scattering and synchrotron radiation microtomography. This is because this method can non-destructively deduce crystalline microstructural information (crystal structure, crystalline phase, crystallographic texture (preferred orientation), microstructure (crystallite size) and crystal lattice strain) of a bulk material over large area (~ 10 cm) with reasonable spatial resolution (~ 100 μm) from the position-dependent Bragg-edge transmission spectra. Furthermore, this method can become a quantitative evaluation technique owing to a Bragg-edge analysis software such as RITS (Rietveld Imaging of Transmission Spectra) [1-3]. For these reasons, it was decided that the first pulsed neutron imaging instruments should be constructed at world-leading pulsed spallation neutron sources, J-PARC MLF in Japan, RAL ISIS in UK, ESS in EU and ORNL SNS in USA. Development activities of the Bragg-edge transmission imaging are typical examples of contribution methodology of a compact accelerator-driven neutron source (CANS). For example, at HUNS (Hokkaido University Neutron Source) in Japan, the Bragg-edge transmission spectroscopy/imaging has been developed [1], and exported to J-PARC MLF in Japan [2,4,5]. At the Bariloche LINAC in Argentina, the Bragg-edge transmission spectroscopy has been developed [6-8], and exported to RAL ISIS in UK [9,10]. Thus, CANS can contribute to the development of new neutron technologies, and export them to worldleading neutron experimental facilities. Here, let us remember two main roles of CANS; not only promotion of such development activity for new neutron technologies (neutron engineering), but also promotion of utilization activity of various neutron experiments (materials science and industrial application). However, actually, the latter’s examples on the Bragg-edge transmission imaging are not so many although it has been already found by the first experimental demonstration at HUNS [1] that the Bragg-edge transmission imaging is also feasible at pulsed CANS. In this paper, we evaluate performances of the Bragg-edge transmission imaging at CANS in detail by using the data of the first quantitative texture/microstructure imaging experiment performed at HUNS [1]. In the discussion, we compare them with data of a high-performance neutron diffraction experiment and the Rietveld analysis performed at a world-leading pulsed neutron experimental facility, J-PARC MLF [11]. Furthermore, we also present a future plan for higher intensity/wavelength-resolution experiment at CANS, based on Monte-Carlo simulation calculation studies. By these items, the feasibility information of the Bragg-edge transmission imaging at CANS can be provided to potential users in detail.
2. Specifications of a pulsed neutron transmission imaging experiment at HUNS In this section, we present the specifications of a pulsed neutron transmission imaging experiment at HUNS, which achieved the first quantitative texture/microstructure imaging with the RITS code [1]. This experiment was carried out under the severest conditions due to high spatial resolution (800 μm), namely, small pixel size causing low neutron counting. In other words, here, we present the specification requirements for quantitative imaging of crystalline structural information at CANS; neutron source, neutron transport and neutron TOF imaging detector. 2.1. Neutron source HUNS is a pulsed cold neutron source based on the s-band 45 MeV electron linear-accelerator installed at Hokkaido University in Japan. Fig. 1 shows the specifications and the schematic view of
255
256
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
TMRA (target-moderator-reflector assembly) of HUNS. For neutron experiments, the Hokkaido University LINAC operates 1 kW (= 33 MeV × 33 μA) of the beam power, and the pulse width of the electron beam is 3 μs with the pulse repetition of 50 Hz. The neutron generation reaction used is the photonuclear reaction caused by bremsstrahlung of electrons. The target material is Pb. The total yield of fast neutrons emitted from the target material is about 1.6×1012 n/s. The high-energy evaporation neutrons are slowing down from the MeV region to the meV region through the neutron moderator. The neutron moderator used at HUNS is usually a coupled type moderator for neutron beam extraction of higher intensity [12]. The main moderator is 18 K solid methane (CH4) of 12 cm × 12 cm × t 5 cm, and the premoderator is polyethylene of 1.5 cm thickness. The neutron pulse width (FWHM of emission time of a certain wavelength neutrons emitted from the main moderator) is about 160 μs for neutrons of the 0.4 nm wavelength. This value is not so narrow because of the high-intensity type coupled moderator. The neutron reflector around the moderator consists of graphite with the dimensions of 80 cm × 80 cm × 80 cm. This volume was decided by a neutronic design based on a Monte-Carlo simulation calculation.
Electron LINAC z 33 MeV㽢33 μA = 1 kW z 3 μs pulse width z 50 Hz pulse repetition Neutron generator z Electron bremsstrahlung z Photonuclear reaction z Pb target z 1.6㽢1012 n/s yield
Polyethylene premoderator (t 1.5 cm) 3RO\HWK\OHQH
Solid methane moderator (12 cm㽢12 cm㽢 t 5 cm)
Neutron moderator z Coupled type (for higher intensity) z 18 K solid CH4 (12 cm㽢12 cm㽢 t 5 cm) z Polyethylene premoderator (t 1.5 cm) z 160 μs pulse FWHM (4ύ neutrons) Neutron reflector z Graphite (80㽢80㽢80 cm3)
10 cm㽢10 cm beam extraction
Lead target Graphite reflector
Fig. 1. Specifications and schematic view of Hokkaido University Neutron Source (HUNS).
2.2. Neutron transport Fig. 2 shows the specifications and the schematic layout of the neutron beam-line geometry for this experiment. We put specimens and a neutron TOF imaging detector at the 6 m position apart from the main moderator surface. Therefore, the neutron wavelength band width was from 0.0 nm to 1.3 nm in case of the 50 Hz pulse repetition. All collimators setup at the beam-line were opened under the condition of the 10 cm × 10 cm square. Therefore, the umbra beam size was 10 cm × 10 cm, and so-called L/D (the collimator ratio) was 60 (± 0.5° beam angular divergence). The neutron flux at the sample/detector
257
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
Sample & Detector position, and Wavelength band width z 6 m from the moderator surface э 1.3 nm wavelength band width Collimators (B4C & Boric-acid resin) z 10 cm㽢10 cm square collimators (10 cm㽢10 cm umbra beam size) Collimator ratio (L/D) z L/D = 60 (= Neutron flight path length 600 cm / Effective moderator size 10 cm) z 㼼0.5㼻beam angular divergence Neutron flux z 8.6㽢103 n/cm2/s @ sample & detector position Wavelength resolution z 2.7% (4ύ neutrons) (= Pulse FWHM 160 μs / Neutron flight time 6067 μs)
Neutron source
Evacuated beam tube Neutron
3.0 m
B4C shield
GEM detector
0.5 m
Specimens
Total flight path length : 6 m Fig. 2. Specifications and schematic layout of the neutron beam-line geometry for this experiment.
position was 8.6×103 n/cm2/s. The neutron wavelength resolution was 2.7% for neutrons of the 0.4 nm wavelength. 2.3. Neutron TOF imaging detector We used a GEM (gas electron multiplier) detector [13] for this experiment. This is because our aim was an experiment with medium spatial resolution (~ 1 mm). The GEM detector is a type of a micro pattern gas detector (MPGD). The GEM detector used in this experiment contained two 10B coated GEM foils and one electron-amplifying GEM foil. The gas used was 70%Ar and 30%CO2 under the pressure of 0.1 MPa. The pixel size (position resolution) due to the readout electrode was 800 μm × 800 μm. The detection area was 9.6 cm × 9.6 cm square composed of 120 × 120 pixels (14,400 positions). The detection efficiency was 15% for neutrons with the wavelength of 0.4 nm. The maximum neutron counting rate depending on the DAQ system was about 106 Hz (1 MHz).
3. Experimental results and comparison with a high-performance neutron powder diffractometer In this experiment, under the experimental condition described in Sec. 2, we measured Bragg-edge transmission spectra at 14,400 positions of rolled/welded α-Fe plates of 6 mm thickness. In this section, firstly, we present and discuss features of Bragg-edge transmission spectra measured at short-pulsed CANS based on the coupled moderator. Furthermore, we present results of quantitative
258
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
texture/microstructure imaging with the Rietveld-type analysis (the RITS code), and also compare them with results of a high-performance neutron powder diffraction experiment with the Rietveld analysis at a world-leading pulsed spallation neutron source. 3.1. Measured Bragg-edge transmission spectrum and its features Fig. 3 (a) shows incident/transmitted neutron spectra, and Fig. 3 (b) shows the neutron transmission spectrum derived from these two spectra (Fig. 3 (a)), measured by one pixel (800 μm × 800 μm area) of the GEM detector. The binning time width of the TOF analysis is 50 μs. The measurement time was 3.3 hours for incident beam, and 5.0 hours for transmitted beam. The measured samples were rolled/welded α-Fe plates of 6 mm thickness. Therefore, a large Bragg-edge caused by {110} diffraction of the bodycentered-cubic (BCC) crystal structure was observed. On the other hand, the data statistics were quite low, about 4% of the statistics error, due to low counts of neutrons, less than 600 counts as shown in Fig. 3 (a). This is because of not only low neutron flux at HUNS (8.6×103 n/cm2/s) but also small pixel size (800 μm × 800 μm) and low detection efficiency (15% for cold neutrons) of the detector.
(a) Raw neutron spectra (1 pixel)
(b) Neutron transmission spectrum (1 pixel) 80%
500 400
Incident neutrons
300
(3.3 hours)
200
Transmitted neutrons
(5.0 hours)
100
70% 60%
Fe {110}
Counts of neutron
600
Neutron transmission
700
50%
40% 30%
0 0.0
0.2
0.4
0.6
0.8
Neutron wavelength / nm
1.0
1.2
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Neutron wavelength / nm
Fig. 3. Neutron spectra measured by one pixel with the TOF bin width of 50 μs. (a) Raw spectra of incident and transmitted neutrons. (b) Neutron transmission spectrum derived from the Fig. 3 (a) spectra.
Furthermore, the broad neutron pulse effect due to the coupled moderator for higher neutron flux is discussed. Fig. 4 shows the data of Bragg-edge transmission spectra for cases of both high and low wavelength resolution, respectively. The data for the case of high wavelength resolution (Δλ/λ = 0.15% for cold neutrons) were measured at J-PARC MLF BL19 “TAKUMI” [14] at the 40 m position from the decoupled-poisoned moderator of JSNS (Japan Spallation Neutron Source) at J-PARC MLF. The data for the case of low wavelength resolution (Δλ/λ = 2.7% for cold neutrons) were measured by this presented experimental setup performed at the 6 m position from the HUNS coupled moderator. The samples were α-Fe plates in both cases. In the low wavelength resolution case, broad Bragg-edges were observed. Therefore, quantitative imaging of crystal lattice strain and crystalline phase is not easy for CANS using a coupled moderator. This is because correct decision of wavelength position of a Bragg-edge (strain analysis) or separation between Bragg-edges (crystalline phase analysis) is not easy. On the other hand, quantitative imaging of texture and crystallite size is not so difficult for such experimental setup because analyzed wavelength regions are far from Bragg-edges (transmission intensities between Bragg-edges). Thus, the data achieved from these experiments indicate that quantitative texture/microstructure imaging can be carried out easily for a coupled moderator at CANS. Incidentally, a methodology to perform quantitative imaging of strain and crystalline phase at CANS is presented in Sec. 4.
259
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
(a) High wavelength resolution case
(b) Low wavelength resolution case 70%
Neutron transmission
Neutron transmission
80% 70%
60% 50% 40% 30%
60%
50%
40% 0.1
0.2 0.3 0.4 Neutron wavelength / nm
0.5
0.1
0.2 0.3 0.4 Neutron wavelength / nm
0.5
Fig. 4. Comparison of Bragg-edge neutron transmission spectra between (a) the high wavelength resolution case (0.15% for 0.4 nm wavelength neutrons) measured at J-PARC MLF BL19 “TAKUMI” and (b) the low wavelength resolution case (2.7% for 0.4 nm wavelength neutrons) measured by this presented experiment.
Energy-selective (wavelength-resolved) neutron radiography also can be performed easily by this experimental setup. Fig. 5 shows wavelength-dependent neutron radiographs measured by this experiment. Two rolled α-Fe plates were set at the upper side in samples, and neutrons were transmitted along the normal direction (ND) of the rolled plates. One welded α-Fe plate was set at the lower side in samples, and neutrons were transmitted along ND of the plate. The welded zone was located along the center-line. The other welded α-Fe plate was set at the middle position in samples, and neutrons were transmitted along the rolling direction (RD) of the plate. Up to the Bragg cut-off wavelength (~ 0.4 nm), lower transmission values are distributed around the welded zone, caused by changes of texture and crystallite size due to the welding processing. It is known that such energy-selective neutron radiography can be related with the SEM-EBSD results complementarily [15]. Thus, at CANS, even if quantitative crystalline microstructural information imaging combined with the Rietveld-type data analysis is difficult, the energy-selective neutron radiography can be performed easily, and can contribute to materials science. 3.2. Quantitative visualization of texture and microstructure with the Rietveld-type data analysis For quantitative imaging of texture and crystallite size, we analyzed all Bragg-edge transmission spectrum data for every pixel by using the RITS code. The results of the imaging were already presented in Ref. [1]. We were successfully able to obtain reasonable images on texture and crystallite size with the reasonable spatial resolution of 800 μm although we carried out this experiment under the condition of quite low neutron counting, as shown in Fig. 3. Owing to the data analysis software “RITS”, we can correctly evaluate crystalline microstructural parameters of the samples at HUNS because the profile (curve) fitting analysis is applicable to low statistics data at CANS. In other words, the RITS code is reactive for low statistics data, and is useful for experiments at CANS. 3.3. Comparison with a neutron diffraction experiment at a huge pulsed spallation neutron source Furthermore, we compared the Bragg imaging results with results obtained by a high-performance neutron powder diffractometer installed at a huge pulsed spallation neutron experimental facility. The pulsed neutron powder diffractometer was iMATERIA at J-PARC MLF BL20 [16]. The Rietveld analysis software was Z-Rietveld developed at J-PARC MLF [17]. During the diffraction experiment, the power
260
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
O = 0.18 nm
O = 0.30 nm 88
0
64
40 -5
88
0
64
40
-5
0 +5 Position x / cm
-5
O = 0.40 nm 88
0
64
40 0 +5 Position x / cm
0
64
40
-5 -5
+5
88
0
64
40
-5 -5
0 +5 Position x / cm
0 +5 Position x / cm
O = 0.60 nm
Cold neutron transmission 䠄%䠅 䠄Wavelength : 0.49 nm 䡚 0.51 nm䠅
Position y / cm
Position y / cm
+5
-5
88
O = 0.50 nm
Cold neutron transmission 䠄%䠅 䠄Wavelength : 0.39 nm 䡚 0.41 nm䠅
-5
+5
0 +5 Position x / cm
Cold neutron transmission 䠄%䠅 䠄Wavelength : 0.59 nm 䡚 0.61 nm䠅
Position y / cm
-5
+5
Low energy neutron transmission 䠄%䠅 䠄Wavelength : 0.34 nm 䡚 0.36 nm䠅
Position y / cm
+5
O = 0.35 nm
Low energy neutron transmission 䠄%䠅 䠄Wavelength : 0.29 nm 䡚 0.31 nm䠅
Position y / cm
Position y / cm
Thermal neutron transmission 䠄%䠅 䠄Wavelength : 0.17 nm 䡚 0.19 nm䠅
+5
88
0
64
40
-5 -5
0 +5 Position x / cm
Fig. 5. Wavelength-dependent neutron radiographs measured by this experiment.
of the J-PARC proton accelerator was 120 kW although the full power is 1 MW, and the neutron yield of JSNS was 104 times higher than that of HUNS. The moderator of BL20 was the decoupled-poisoned 20 K supercritical 100% para-H2 moderator. 3Qc supermirror guide tube of 14 m length was installed. The distance from the moderator to the sample was 26.5 m, and the distances from the sample to the detectors were 2.0 ~ 2.3 m. The high-resolution backscattering detector bank was used. The resolution of crystal lattice plane spacing was 0.16%. The measurement time was 30 minutes for one irradiation. One irradiation can investigate only one position in a sample. In this diffraction experiment, the same samples as the Bragg imaging experiment at HUNS were measured, which is already presented in Sec. 3.1 and 3.2. Fig. 6 shows results of the comparison between the Bragg-edge transmission imaging with the RITS code at HUNS and the diffraction with the Z-Rietveld code at J-PARC MLF BL20 “iMATERIA”. Both of the results (texture degree and crystallite size) for any samples’ position/direction were corresponded within the analysis error. Correctly, only for crystallite size, 1.55 times larger values of the diffraction analysis can correspond with the results of the Bragg-edge analysis. We guess this difference is caused by neutron wavelength resolution. Higher resolution can sharpen a diffraction peak, and can weaken the extinction effect. For the crystallite size analysis, both codes use the extinction effect caused by rediffraction of neutrons inside a crystallite. Therefore, the high-resolution diffraction experiment evaluated smaller extinction effect and finer crystallites than the low-resolution Bragg imaging experiment. Thus, the Bragg-edge transmission imaging at CANS can give the reasonable data with high precision, which is consistent with a high-performance neutron powder diffractometer installed at a world-leading pulsed spallation neutron source. In other words, the Bragg-edge transmission imaging is expected to be one of the most efficient crystallographic/metallographic analysis tools for pulsed CANS.
261
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
1.0
(b) Crystallite size
Bragg imaging Diffraction
0.8
0.6
Weld/ND Weld/RD Roll/ND Roll/RD
0.4 0.2 0.0
6.0
Crystallite size / μm
March-Dollase coefficient
(a) Degree of crystallographic anisotropy (March-Dollase coefficient)
5.0
Roll/RD
4.0
Bragg imaging Diffraction×1.55 Diffraction
Roll/ND
3.0
Weld/RD Weld/ND
2.0 1.0 0.0
0
1 2 3 4 Evaluated position/direction
5
0
1 2 3 4 Evaluated position/direction
5
Fig. 6. Results of the comparison between the Bragg-edge transmission imaging with RITS at HUNS and the diffraction with ZRietveld at J-PARC. (a) Degree of crystallographic anisotropy, indicated by the March-Dollase coefficient. (b) Crystallite size.
4. Future prospect toward quantitative imaging of strain and crystalline phase at CANS Due to the low wavelength resolution of a coupled moderator for higher low-energy neutron yield at CANS, quantitative analysis/imaging of strain and crystalline phase using the Bragg-edge transmission imaging is not easy for CANS. On the other hand, in case of a decoupled moderator use, the higher wavelength resolution can be achieved, but the neutron flux becomes weaker than the coupled moderator case. Therefore, we considered an experimental setup using a supermirror guide tube for efficient neutron beam transport, combined with the decoupled moderator. Hereafter, we discuss effects of such experimental setup at CANS by using a Monte-Carlo neutron transport simulation code, McStas [18]. Fig. 7 shows a schematic layout for the simulation. This geometry is a reasonable beam-line design at HUNS. The sample/detector position was the 6 m position apart from a decoupled hydrogen moderator. At the just center position of the whole beam-line, a 3.65Qc supermirror guide tube of 4 m length was set. The open area of the tube was 10 cm × 10 cm square. In the simulation, we evaluated wavelength resolution (sharpness of a Bragg-edge) due to the decoupled moderator, neutron flux complement due to the supermirror guide tube, and increase of beam angular divergence due to the supermirror guide tube. Fig. 8 (a) shows sharpness of a Bragg-edge which will be measured by the experimental setup as shown in Fig. 7. The result is presented together with the Δλ/λ = 2.7% data of the coupled CH4 moderator setup at HUNS (the neutron flight path length was 6 m), and the Δλ/λ = 0.35% data of the decoupled supercritical 100% para-H2 moderator setup at J-PARC MLF BL10 “NOBORU” [19] (the neutron flight path length was 14 m). According to Fig. 8 (a), at CANS, a sharp Bragg-edge can be observed owing to the decoupled moderator, and the wavelength resolution of 0.8% can be achieved. This value is not so bad for the strain/phase analysis of the Bragg-edge transmission spectroscopy. Fig. 8 (b) shows a result of the neutron flux recovery owing to the supermirror guide tube. In case of only the decoupled hydrogen moderator, the neutron flux becomes quite weak, about 10% intensity of the coupled hydrogen moderator. In case of the decoupled moderator setup combined with the supermirror guide tube, the neutron flux can increase up to 150% intensity of the coupled moderator. The better intensity complement is expected in the methane moderator case than the hydrogen moderator case. Finally, Table 1 shows change of the beam angular divergence before/after setting of the supermirror guide tube, with information of its corresponding L/D value. After the set of the supermirror guide tube, the beam angular divergence increases, but is not so bad. This is because the 1 mm spatial resolution,
262
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
reasonable value for quantitative crystalline microstructural information imaging using the Bragg-edge transmission spectroscopy, can be achieved for a several-mm thickness sample directly coupled with an imaging detector (no distance between the sample and the detector). Thus, owing to the decoupled moderator setup combined with the supermirror guide tube, quantitative imaging of not only texture and microstructure but also strain and crystalline phase will be reasonably performed at pulsed CANS. Decoupler
3.65Qc supermirror guide tube
Moderator
Sample Imaging detector
Reflector
4m 6m
Fig. 7. Schematic layout of the McStas simulation for a higher wavelength resolution experiment at HUNS/CANS.
Total cross section / a.u.
0.8
Fe {110} 0.7
0.6 0.5 0.4
Δλ/λ = 0.35% (J-PARC MLF BL10)
(b) Gain of neutron intensity
Δλ/λ (6 m from decoupled HUNS) ~ 0.8% Δλ/λ = 2.7% (HUNS)
0.3
0.38
0.40 0.42 0.44 0.46 Neutron wavelength / nm
0.48
Flux ratio to coupled HUNS
(a) Sharpness of Bragg edge
2.5 2.0
1.5
Decoupled moderator w/ guide tube
1.0
Coupled moderator
0.5
Decoupled moderator 0.0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Neutron wavelength / nm
Fig. 8. (a) Sharpness of a Bragg-edge due to the decoupled moderator setup. (b) Gain of the decoupled hydrogen moderator setup combined with/without a supermirror guide tube. The gain means the ratio to intensity of the coupled hydrogen moderator. Table 1. Comparison of beam angular divergence before/after setting of the supermirror guide tube.
Without supermirror guide tube With supermirror guide tube
Beam angular divergence ± 0.5° (L/D ~ 60) ± 2.4° (L/D ~ 12.5)
5. Conclusion We evaluated and discussed performances of the Bragg-edge transmission imaging at a compact accelerator-driven short-pulsed neutron source in order to indicate the detailed feasibility of this new useful technique at a popular-priced facility for industrial use. A coupled moderator is usually used to gain higher beam intensity. Therefore, quantitative imaging of strain and crystalline phase is not easy due to broad Bragg-edges. For this problem, it was found that a high wavelength resolution experimental setup
Hirotaka Sato et al. / Physics Procedia 60 (2014) 254 – 263
using a decoupled moderator connected to a supermirror guide tube is effective, according to the MonteCarlo simulation calculation based neutronic studies. Even if the coupled moderator is used, quantitative imaging of crystallographic texture and microstructure (crystallite size) can be achieved. The important matter to analyze low statistics data is to use the data analysis software that can perform the profile (curve) fitting toward experimental data, like the Rietveld-type data analysis software, RITS. Furthermore, such software can provide quantitative data that are consistent with results of a high-performance neutron diffractometer at a world-leading pulsed spallation neutron source. Thus, the Bragg-edge transmission imaging is expected to be one of the most efficient crystallographic analysis tools for a compact accelerator-driven short-pulsed neutron source. Acknowledgements This work was partially supported by Grant-in-Aid for Scientific Research (S) from Japan Society for the Promotion of Science (No. 23226018). References [1] H. Sato, T. Kamiyama and Y. Kiyanagi, Mater. Trans. 52 (2011) 1294-1302. [2] Y. Kiyanagi, H. Sato, T. Kamiyama and T. Shinohara, J. Phys. Conf. Ser. 340 (2012) 012010. [3] H. Sato, T. Shinohara, R. Kiyanagi, K. Aizawa, M. Ooi, M. Harada, K. Oikawa, F. Maekawa, K. Iwase, T. Kamiyama and Y. Kiyanagi, Phys. Proc. 43 (2013) 186-195. [4] Y. Kiyanagi, T. Kamiyama, H. Sato, T. Shinohara, T. Kai, K. Aizawa, M. Arai, M. Harada, K. Sakai, K. Oikawa, M. Ohi, F. Maekawa, T. Sakai, M. Matsubayashi, M. Segawa and M. Kureta, Nucl. Instrum. Methods A 651 (2011) 16-20. [5] Y. Kiyanagi, T. Shinohara, T. Kai, T. Kamiyama, H. Sato, K. Kino, K. Aizawa, M. Arai, M. Harada, K. Sakai, K. Oikawa, M. Ooi, F. Maekawa, H. Iikura, T. Sakai, M. Matsubayashi, M. Segawa and M. Kureta, Phys. Proc. 43 (2013) 92-99. [6] F. Kropff and J. R. Granada, Unpublished Report CAB-1977, Institute Balseiro, Bariloche (1977). [7] J. R. Granada, J. R. Santisteban, J. Dawidowski and R. E. Mayer, Proceedings of 9th International Topical Meeting on Nuclear Research Applications and Utilization of Accelerators (AccApp09), AP/IE-02, International Atomic Energy Agency, Vienna (2010). [8] S. Petriw, J. Dawidowski and J. Santisteban, J. Nucl. Mater. 396 (2010) 181-188. [9] J. R. Santisteban, L. Edwards, M. E. Fitzpatrick, A. Steuwer, P. J. Withers, M. R. Daymond, M. W. Johnson, N. Rhodes and E. M. Schooneveld, Nucl. Instrum. Methods A 481 (2002) 765-768. [10] W. Kockelmann, S. Y. Zhang, J. F. Kelleher, J. B. Nightingale, G. Burca and J. A. James, Phys. Proc. 43 (2013) 100-110. [11] H. Sato, T. Kamiyama, K. Iwase, T. Ishigaki and Y. Kiyanagi, Nucl. Instrum. Methods A 651 (2011) 216-220. [12] Y. Kiyanagi, N. Watanabe and H. Iwasa, Nucl. Instrum. Methods A 312 (1992) 561-570. [13] S. Uno, T. Uchida, M. Sekimoto, T. Murakami, K. Miyama, M. Shoji, E. Nakano and T. Koike, Phys. Proc. 37 (2012) 600-605. [14] S. Harjo, T. Ito, K. Aizawa, H. Arima, J. Abe, A. Moriai, T. Iwahashi and T. Kamiyama, Mater. Sci. Forum 681 (2011) 443448. [15] E. H. Lehmann, S. Peetermans, L. Josic, H. Leber and H. van Swygenhoven, Nucl. Instrum. Methods A 735 (2014) 102-109. [16] T. Ishigaki, A. Hoshikawa, M. Yonemura, T. Morishima, T. Kamiyama, R. Oishi, K. Aizawa, T. Sakuma, Y. Tomota, M. Arai, M. Hayashi, K. Ebata, Y. Takano, K. Komatsuzaki, H. Asano, Y. Takano and T. Kasao, Nucl. Instrum. Methods A 600 (2009) 189-191. [17] R. Oishi, M. Yonemura, Y. Nishimaki, S. Torii, A. Hoshikawa, T. Ishigaki, T. Morishima, K. Mori and T. Kamiyama, Nucl. Instrum. Methods A 600 (2009) 94-96. [18] K. Lefmann and K. Nielsen, Neutron News 10 (1999) 20-23. [19] F. Maekawa, K. Oikawa, M. Harada, T. Kai, S. Meigo, Y. Kasugai, M. Ooi, K. Sakai, M. Teshigawara, S. Hasegawa, Y. Ikeda and N. Watanabe, Nucl. Instrum. Methods A 600 (2009) 335-337.
263