Neutron scattering instrumentation at low power reactors for science,engineering and education

Abstract

We demonstrate that a suite of neutron scattering instruments can be installed on low power reactors from e.g. 0.1 to 1 MW thermal power, including a powder diffractometer, a Laue diffractometer, a neutron imaging station, and small angle scattering machine. For each of these instruments, we estimate the performance using Monte-Carlo ray-tracing simulations with McStas, propose a cost effective design, and a set of applications covering science, engineering and education. In addition, we suggest a few shielding solutions to minimize the background level.

Keywords

Neutron scattering research reactors instrumentation Monte-Carlo simulation McStas powder diffraction imaging Laue diffraction small angle scattering

1. Introduction

There are currently about 250 operational research reactors in the world [13]. Today, most facilities provide some of neutron activation analysis, radioisotope production, material testing and transmutation from neutron irradiation, as well as a nursery for education and training. However, only about 30 of these reactors are equipped with neutron scattering instrumentation. Restricting the research reactors to those providing at least 100 kW thermal power, about 70–80 additional facilities could advantageously equip their neutron beam ports with neutron scattering instruments to enhance the utilisation of the research reactor use.

As the neutron flux obtainable at e.g. 100 kW to few MW thermal reactors is limited, we shall mostly focus our study on structure characterisation instruments (elastic neutron scattering). The signal from inelastic neutron scattering instruments (probing atom motions) is usually much weaker than that probing atom arrangements (crystallography and diffraction), and will therefore not be discussed here.

In the past, initiatives to install such instrumentation has been reported e.g. at the TRIGA reactors in Pavia, Italy [4], Vienna, Austria [3], Ljubjana, Slovenia [26], Bangi, Malaysia [28]. Such MW-class research reactors have been installed in many countries since 1960, for an installation cost of about 3 M$/unit.

The main purpose of this study is to demonstrate that existing low power research reactors can benefit from neutron scattering instruments at a reduced installation and maintenance cost. In this way, local communities of neutron scattering users can emerge around these facilities, in order to train young scientists and disseminate the possibility to use these methodologies. Such new users can then join larger neutron research facilities when needing more advanced instrumental configurations and higher neutron fluxes.

In the following, we shall detail the characteristics of a typical thermal power research reactor and derive model parameters for further simulations. Then, we shall demonstrate that simple yet powerful instruments can be installed at these facilities. Namely, we propose a low resolution powder diffractometer, a Laue diffractometer, a neutron imaging station, and a small to medium angle scattering diffractometer. For each of these instruments, we shall estimate its flux, resolution and typical acquisition signal. Suggestions for installation will be made, using commercial parts. Last, we shall discuss background issues, and simple solutions for shielding.

2. Reactor model

Most existing medium power research reactors are built around a nuclear core using low-enriched uranium fuel elements, surrounded by a reflector (usually graphite or beryllium), in a light water tank acting as coolant, reflector and biological shielding. The reactor core is in thermal equilibrium with the water moderator, which defines the surrounding neutron bath temperature. Often, the reactor core has vertical irradiation loops to produce radioisotopes, and a number of horizontal neutron beam ports. These beam ports can be radial (directly pointing at the core, with or without reflector inlet), or tangential (pointing at the core reflector), as shown in Fig. 1.

Fig. 1.

Cut away view of a TRIGA Mark II type reactor, with a typical thermal power of 1 MW. The four beam ports provide a neutron thermal flux around 2 × 10⁸ n/cm²/s with a maximum brilliance around $λ = 1.15 Å$ (750 kW thermal power).

As an example we shall focus on the Puspati TRIGA reactor installed in Kajang, Malaysia, with a nominal thermal power of 1 MW. The core diameter is 45 cm, the graphite reflector diameter is 106 cm, and the water tank diameter is 2 m. The in-core thermal neutron flux has been measured as 2.3 × 10¹² n/cm²/s/MW_th [1]. Using a source brilliance parameter $I = 3 \times 10^{11} {n/cm}^{2} {/s/sr/Å/MW}_{th}$ for a black-body temperature $T = 330 K$ , we can simulate the reactor neutron flux in the thermal range from e.g. 2 meV up to 300 meV using McStas [16,32]. The epithermal contribution from the core is not modelled. The maximum brilliance is obtained at $λ = 1.15 Angs$ (62 meV) with a peak brilliance $ϕ_{max} = 5 \times 10^{12} {n/cm}^{2} /s/sr/Å$ at a thermal power of 750 kW (Fig. 2).

Fig. 2.

Brilliance curve at a thermal power of 750 kW, computed at the outlet of a radial neutron beam port. The epithermal contribution from the core is not modelled. The maximum brilliance is at neutron wavelength $λ = 1.15 Å$ . This curve can be scaled to the actual power of other thermal research reactors to get an estimate of their brilliance.

At a radial beam port outlet ( $ϕ = 15.4 cm$ ), which is typically located at about 3.3 m from the core centre, the wavelength integrated flux is computed as 2 × 10⁸ n/cm²/s (Fig. 3) and the maximum flux is computed as 1.8 × 10⁸ n/cm²/s/Å at $λ = 1.15 Å$ for a thermal power of 750 kW. The corresponding beam port neutron fluence rate is 240 n/cm²/s/W_th in agreement with [30]. The beam divergence is found as 2.1° (full width), as shown in Fig. 4. The actual flux depends slightly on the beam port geometry. The most intense beam ports (radial piercing) benefit from a higher flux (e.g. +50%) but as well from a larger gamma and epithermal neutron background. On the other side, tangential beam ports have a lower neutron flux (e.g. −50%), but a reduced gamma contamination.

Fig. 3.

Flux at a thermal power of 750 kW, computed at the outlet of a radial neutron beam port. The maximum flux is at neutron wavelength $λ = 1.15 Å$ . The integrated flux is found around 2 × 10⁸ n/cm²/s for a thermal power of 750 kW, at 3.3 m from the core centre.

Fig. 4.

Beam divergence, at the outlet of a radial neutron beam port for a thermal power of 750 kW, at 3.3 m from the core centre.

The installation of a super-mirror coating along the in-pile collimator has been tested, but as expected on such a short guide distance, it does not bring any benefit in terms of intensity at the beam port.

3. Overview of proposed instruments

Neutrons, due to their neutral charge, have a high penetrating power, which makes them fully suited to study bulk materials, as opposed to e.g. compact X-ray diffractometers and light spectrometry which mostly probe the near surface material behaviour.

Due to limited flux at low power thermal research reactors, we focus our study on imaging and diffraction techniques. The wavelength range for neutrons delivered by such a facility determines the materials that can be studied. As the maximum brightness is obtained around 1.1 Å neutron wavelength, diffraction is the technique of choice to make best use of the neutron flux. However, a reasonable yet limited flux is usable up to 5 Å neutron wavelength, which opens the possibility to install a small-to-medium angle scattering instrument.

The criteria for selecting instrument design is cost and easy installation using standard commercial solutions as most of these facilities do not possess technical laboratories able to manufacture monochromators, detectors, motorised stages and goniometers, collimators.

We propose to install a powder diffractometer (to e.g. study the structure of solid-state materials), a Laue diffractometer (to e.g. study the structure of small proteins), a small-medium angle instrument (to e.g. study the structure of small polymer chains), and a radiography station (to e.g. image the internal parts of macroscopic objects). All of these instruments are presented with a simple and cost-effective configuration and are made of a limited set of commercial interchangeable parts. An overview of the installed instruments as modelled with McStas [16,32] is shown in Fig. 5. The rationale for all instruments design is to use an as compact as possible configuration. This is clearly at the cost of a coarse resolution but provide versatile, high acquisition rate instruments even on low power research reactors, as demonstrated below.

Fig. 5.

Overview of the four proposed neutron scattering instruments installable at a low power research reactor, modelled with McStas. The four beam ports are equipped with a small angle scattering instrument, a powder diffractometer, a Laue diffractometer and an imaging station. The axes are indicated in meters.

The radial beam ports provide the highest neutron flux, at the cost of a significantly stronger gamma and epithermal neutron background. However, using monochromators to deviate the beam, the actual sample and detectors remain off axis. Such radial beam ports are then suited for the installation of the powder diffractometer, and the small-angle instrument. In order to limit the total investment, these two instruments may share the same beam-port, to minimize the cost of the shielding.

The tangential beam ports should be preferred for white beam instruments, such as the Laue diffractometer and the imaging station. Indeed, the gamma and epithermal neutron background should be reduced as much as possible to make use of cost-effective scintillation detectors. The CCD detector modules can also be shared between the instruments, when installed as mountable racks.

In the following, we estimate the performance of the four proposed instruments using the neutron ray-tracing Monte-Carlo software McStas [16,32]. A global optimisation of instrumental parameters has been carried out using iFit [8] to find the best flux/resolution ratio configuration for each instrument.

4. Powder diffractometer

The powder diffractometer is a versatile instrument that aims at measuring the structure of solid-state materials. It can also be used to study liquids and amorphous systems. The measurement is mostly based upon the Bragg law $\begin{matrix} 2 d sin θ = n λ, \end{matrix}$ where λ is the incident neutron wavelength, n is the order of the scattering, $2 θ$ is the scattering angle at the sample and d is the distance between atomic planes. In a powder or liquid, for a given incoming neutron wavelength and d-spacing, the scattering pattern is detected as a set of Debye–Scherrer cones of full opening $4 θ$ which intersection with detectors show-up as ring arcs. In addition, in a powder, the intensity of the diffraction rings allows to determine the type and position of the atoms in the crystallographic lattice cell of the material.

In practice, when installed on a thermal research reactor as described above, the neutron wavelength λ is between 1 and 3 Å. The scattering angle is often in the range 10–150° which then allows to probe materials with inter-atomic distances of the order of a few Å specially suited to study most ceramics and small crystal structures. This would for instance be the case of geological materials for the mining industry and jewellery, metallurgy, ceramics, phase transitions, and of course for teaching crystallography.

The neutron beam from a radial beam port is collimated to the size of the down-stream monochromator. An optional 10 cm thick single crystal sapphire filter [19] can be installed immediately after the main beam shutter, to remove the fast neutron reflection orders at the monochromator for wavelength $λ < 0.2 Å$ . The large monochromator, positioned at 3.3 m from the core-centre, should advantageously be focusing with a fixed curvature optimised for the nominal diffractometer wavelength configuration, e.g. $λ = 2.36 Å$ . The monochromator must be mounted on a motorised 2-axes goniometer, to ease the alignment of the beam towards the sample location. We have envisaged two distinct diffractometer configurations at $λ = 1.53$ and 2.36 Å incident neutron wavelengths, assuming a pyrolytic graphite monochromator is used on the (002) reflection. These configurations allow to optionally use a highly oriented pyrolytic graphite (HOPG) filter [2] to remove most of the 2nd and 3rd reflection orders, typically 5 cm thick. Such a filter would be installed after the monochromator, immediately after the secondary beam shutter located on the monochromator shielding block. A slit should be installed to collimate the beam before the sample location, down to the size of the sample. The detector is modelled as an ideal cylindrical detector covering angles from −20° to 170° with a height of 20 cm, after removing the direct transmitted beam. In addition, a set of 4 CCD panels cover an equivalent angular range, with a nominal efficiency of 9.2%, assuming ZnS(Ag)–⁶LiF scintillator membranes. More details on the instrument parts are given in a dedicated section below. An overview of the powder diffractometer is shown in Fig. 6. The McStas diffractometer model uses the curved monochromator and the powder sample component [33].

Fig. 6.

Overview of the powder diffractometer model generated by McStas. The neutron beam comes from the left side through the collimator, is filtered by the sapphire filter, reflected by the monochromator, passes through the highly oriented pyrolytic graphite filter. The beam is reduced to the size of the sample by a slit, and is then scattered by the sample. The scattered neutrons are collected by a set of CCD panels arranged in a cylinder. The main parameters are detailed in Table 1.

In order to determine an optimal set of parameters for the diffractometer, we have performed a global optimisation of the instrument parameters, for each of these configurations.

The optimal configuration is found using iFit [8], selecting the scattered flux on the final detector as global optimisation criteria. The particle swarm optimiser [14] is used to search the instrument parameters that provide the highest flux at the detectors. The optimisation procedure converges in about 250 iterations. The results are shown in Table 1, and correspond with a very compact geometry.

Table 1

Powder diffractometer optimal parameters with a PG(002) monochromator operated at $λ = 1.53 Å$ and $λ = 2.36 Å$ incident neutron wavelength. The final flux at the detector is an estimate which depends on the detector technology, the sample-detector distance and the sample. The reactor power is set to 750 kW

Powder diffractometer
Wavelength [Å]	$λ = 1.53 Å$	$λ = 2.36 Å$
Beam port	Radial	Radial
Beam port diaphgram [width × height]	10.1 × 15.2 cm²	11.8 × 15.0 cm²
Single crystal sapphire filter at beam port	Yes, 10 cm thick	Yes, 10 cm thick
Monochromator size [width × height]	14.8 × 14.3 cm²	12.7 × 13.3 cm²
Monochromator curvature [horz., vert.]	$R_{H} = 3.8 m$ , $R_{V} = 0.30 m$	$R_{H} = 2.75 m$ , $R_{V} = 0.43 m$
Core-Monochromator distance	$L_{1} = 3.3 m$	$L_{1} = 3.3 m$
Monochromator mosaicity	$η = 50 arcmin PG (002)$	$η = 44 arcmin PG (002)$
Take-off [°]	$2 θ = 26.36 °$	$2 θ = 41.19 °$
Monochromator-sample distance	$L_{2} = 0.72 m$	$L_{2} = 0.75 m$
HOPG filter after monochromator	Yes, 5 cm thick	Yes, 5 cm thick
Beam size at sample [width × height, cm]	2.2 × 4 cm²	1.6 × 2 cm²
Sample-detector distance	$L_{3} = 0.40 m$	$L_{3} = 0.48 m$
Flux at sample [n/cm²/s]	12.5 × 10⁶	5 × 10⁶
Maximum intensity at detector with NAC sample [n/s]	4.1 × 10⁵ (ideal)	1.35 × 10⁵ (ideal)
Detector, ZnS(Ag)–⁶LiF or low pressure ³He panels	4 panels, 20 × 20 cm²	4 panels, 20 × 20 cm²

We present in Fig. 7 a typical diffractogram from a $ϕ 1 cm$ Na₂Ca₃Al₂F₁₄ sample rod, as measured in a simulated 5 minute exposure, and the $λ = 2.36 Å$ configuration. The diffractogram clearly shows the expected diffraction rings.

Fig. 7.

Simulated diffractogram from a Na₂Ca₃Al₂F₁₄ powder sample obtained on an ideal cylindrical detector (top), and on 4 lower sensitivity ZnS(Ag)–⁶LiF CCD panels (bottom), for a 5 minute exposure with the $λ = 2.36 Å$ configuration. The incoming beam is filtered with HOPG and sapphire blocks. Intensity levels range from black to white.

It can be seen that the optimised flux at the sample position compares with that currently available at world-class powder diffractometers, at the cost of a coarser angular resolution. With proper shielding (see dedicated section below) and filters, the signal to noise ratio can be optimised to enhance the diffractogram contrast. In practice, such an instrument can then be used to study most of the powder and liquid/amorphous materials to produce scientific and engineering results.

Fig. 8.

Simulated diffractogram from a Na₂Ca₃Al₂F₁₄ powder sample with a $λ = 2.52 Å$ configuration, compared with a measured diffractogram on D1B@ILL with $L_{3} = 1.5 m$ . The simulated incoming beam is filtered with HOPG and sapphire blocks.

In order to estimate the capabilities of such a powder diffractometer in terms of angular resolution, we compare the integrated diffractogram for the compact configuration detailed in Table 1, transposed to the nominal D1B neutron wavelength $λ = 2.52 Å$ with that obtained on the D1B instrument at the ILL, using $L_{3} = 1.5 m$ and without HOPG filter. The D1B diffractometer is installed on a super-mirror coated thermal neutron guide using a PG(002) monochromator which provides 6.5 × 10⁶ n/cm²/s at the sample position. In this respect, the two instruments have similar incoming fluxes at the sample position. However, as can be seen in Fig. 8, the overall angular resolution of the proposed powder diffractometer is much coarser, as expected. However, we still believe that a proper Rietveld refinement can be achieved with such simulated results to determine the structure of powders with lattice parameters up to 10–15 Å.

Once built and operational, the powder diffractometer can be upgraded for more targeted usage. This includes a strain scanning option with the installation of two narrow slits, typically 1 mm, immediately before and after the sample location to define a small gauge volume. The sample is then iteratively translated e.g. along the incident beam, and the intensity and shape of a single sharp diffraction peak, usually around 90° scattering angle, is recorded to measure the changes in lattice spacing related to the material strain and local deformations. This allows for instance to probe non destructively internal constraints in mechanical parts from which potential cracks may grow. The same set-up also allows to record the full diffractogram from the small gauge volume inside the sample, to e.g. probe local phase transformations. When the material exhibits a microstructure preferred orientation, such as in laminated alloys, the texture will show-up as intensity variations along diffraction rings for each atomic plane reflection. For better performances, and lower sensitivity to the gamma radiation background, the detector may be upgraded to a low pressure ³He detector.

The measured data can be reduced by LAMP [21] and analysed using FullProf [22] and GSAS [15,29].

As will be seen below, the diffractometer can advantageously share its monochromator with the small angle scattering machine, to minimise the cost of the neutron optics and shielding.

5. Imaging/radiography station

A neutron imaging instrument is one of the simplest instruments to build on a low power research reactor. It mostly consists in a fully opened beam tube, a slit $ϕ D$ and a sapphire filter to remove fast neutrons at the beam port outlet, a set of diaphragms to adapt the beam size to that of the sample, the sample to measure and a CCD panel in transmission.

The principle of the neutron imaging is that the transmission through an object is given as: $\begin{matrix} T = e^{- ρ σ (E i) d}, \end{matrix}$ where ρ is the scatterer density, $σ (E_{i})$ is the incident neutron energy dependent total cross section $\begin{matrix} σ (E_{i}) = σ_{abs} \sqrt{(25.3 / E_{i})} + σ_{inc} + σ_{coh} (E_{i}) \end{matrix}$ for thermal neutrons, and d is the object thickness along the beam. The total cross section includes the absorption, as well as the coherent and incoherent scattering contributions. The coherent cross section $σ_{coh} (E_{i})$ presents so called Bragg-edges which are characteristic of the material, and show up as drops in the scattering cross section every time the Bragg condition can not be fulfilled when decreasing the incident neutron energy.

A neutron imaging station can be used for archaeology and cultural heritage imaging to probe non-destructively the internal structure of invaluable objects. Also, it can be used to locate cracks and defects in materials for engineering purposes.

One of the limitations of this radiography station originates from the reactor gamma and epi-thermal neutrons. This constant background level can be reduced by selecting a tangential neutron beam port, installing a proper shielding at the beam port outlet, a 10 cm thick sapphire filter [19], as well as heavy concrete blocks around the imaging station. However, the coherent and incoherent scattering through the material may create a background signal which limits the contrast and highly depends on the material. Many archaeology artefacts are made of ceramics and plaster, potentially containing a significant fraction of hydrogenated or partially wet materials whose scattering power is high. This scattering background can not be reduced, except by e.g. deuterating the sample.

The spatial resolution in the sample is determined by the divergence distribution. When the sample is located immediately after the beam port outlet, every pixel at the detector may receive signal from any point on the neutron tube entrance cap, as there is no collimation to limit the beam divergence. The spatial resolution on the transmitted images is coarse, with an $L / D$ determined by the in-pile collimator length and the beam port section, that is $L = 3.3 m$ and $D = 15.4 cm$ in a TRIGA reactor. The flux, without any collimation, at the sample location is about 10⁸ n/cm²/s for a 750 kW reactor power. The resolution can be highly improved by placing a $ϕ 1 {cm}^{2}$ slit at the beam port to actually increase the $L / D$ ratio, followed by the sapphire filter. In such a set-up, the in-pile collimator should be wide open. The sample must then be moved backwards at typically $L_{1} = 2 – 3 m$ away to be able to illuminate 10–15 cm objects from the slit acting as a virtual source. The flux at the sample is then reduced to about 10⁶ n/cm²/s at the sample location for a 750 kW reactor power, but the angular accuracy is increased to $L / D \sim 330$ , allowing to resolve sub-mm details. Adapting the slit diameter D allows to control the required image resolution, with the collateral flux reduction. Allowing more space $L_{1}$ down stream allows to illuminate larger samples.

We have simulated the imaging station using a $ϕ 1 cm$ slit as virtual source, located at the beam port outlet, 3 m away from the reactor core through the neutron beam port collimator. The slit provides a divergent beam illuminating the sample, located at a distance $L_{1} = 3.4 m$ downstream. The CCD panel used in transmission geometry immediately after the sample has a nominal efficiency of 9.2%, assuming ZnS(Ag)–⁶LiF scintillator membranes. More details on the instrument parts are given in a dedicated section below. A schematic representation of this set-up is shown in Fig. 9. The model makes use of the McStas powder sample component [33] and its ability to handle sample geometries from CAD software [9]. The parameters and performance of the imaging station are summarized in Table 2. The resulting image is presented in Fig. 10 for a transmission through a 10 cm mechanical socket, as well as through a box containing an ancient artefact. From the radiography images, the spatial resolution in the sample with this set-up is measured as about 0.3 mm.

Fig. 9.

Overview of the neutron radiography model generated by McStas. The neutron beam comes from the left side through the collimator (length 3 m), and is filtered by a slit $ϕ D$ and a sapphire single crystal. The beam is reduced to the size of the sample immediately before it, and is then both scattered and transmitted by the sample. The neutrons are collected by a CCD panel positioned in transmission immediately after the sample. The heavy concrete shielding is represented as a 2 m block. The spatial resolution is quantified by the $L_{1} / D$ ratio.

Table 2

Neutron radiography station parameters. The reactor power is set to 750 kW

Neutron radiography
Beam port	Tangential, full opening
Slit at beam port	$ϕ D = 1 {cm}^{2}$ at 3 m from core
Single crystal sapphire filter at beam port	Yes, 10 cm thick
Slit-sample distance	$L_{1} = 3.4 m$
Maximum sample size	15 cm
Flux at sample [n/cm²/s]	10⁶
Detector, ZnS(Ag)–⁶LiF	20 × 20 cm², at $L_{2} = 20 cm$ from sample
Intensity at detector [n/s]	2 × 10⁷ (ideal)
L/D	330
Typical spatial resolution	300 µm for a 10 cm object

Fig. 10.

Simulated neutron radiography obtained at a tangential beam port. The initial sample geometry (left) and the transmitted image (centre) are shown. The right image shows the neutron radiography from an aluminium box containing an ancient artefact. The slit at the beam port outlet is $ϕ 1 cm$ followed by a sapphire filter. The sample is 3.4 m away, and the CCD panel is 20 × 20 cm² located immediately after the sample. Intensity levels range from white to black (negative image).

Adding a goniometer to rotate the sample opens the possibility to perform computed tomography (CT) scans, and to reconstruct the sample volumetric geometry from individual rotated exposures. Current software that may be used for data processing are NiftiRec [20], SPIERS [27], IMOD [17], PITRE [6].

The Bragg-edge option [24], which consists in recording the radiography image while scanning a narrow wavelength band, e.g. from a velocity selector, would require a higher incoming neutron flux, and is probably out of scope here.

As will be seen below, the imaging station can advantageously share its location with the Laue diffraction instrument, to minimise the cost of the neutron optics and shielding.

6. Laue diffractometer

The neutron Laue diffraction method in transmission geometry allows to record the Bragg scattering on atomic planes in a single crystal material, using a white beam incoming neutron beam. Each spot recorded on the detector obeys the Bragg law $\begin{matrix} 2 d sin θ = n λ, \end{matrix}$ where n is the scattering order. However, as the neutron wavelength λ covers a wide range, most atomic plane spacing d may find a suitable scattering angle $2 θ$ . The direction for the scattering is, as opposed to powders and liquids where it appears as Debye–Scherrer cones, perfectly defined in a single crystal along discrete axes, which appear as spots on the detector. A structure refinement can then be performed using the detector image to determine the atom type and position in the lattice unit cell.

A Laue diffraction instrument for single crystal samples has the same geometry as the imaging station. The only difference is the small size of the sample and an additional slit before it. Also, a set of CCD panels must be arranged in transmission and scattering geometry, similarly to that used in the proposed powder diffractometer above. Consequently, the imaging and the Laue instruments may share the same beam port. A high precision goniometer should be installed to hold the sample, so that it can be oriented along high symmetry axes, to ease the structure and space group identification.

Such a diffractometer can be used to study both organic, e.g. small proteins, ordered liquid crystal phases, and inorganic single crystals, e.g. ceramics, geological and jewellery crystals. However, when studying proteins, which are mostly hydrogenated molecules, a large incoherent background is superposed with the Laue diffraction pattern, reducing the contrast. Deuterated proteins should then be preferred. Reciprocal space surveys can also be achieved, studying the effect of external parameters and composition. Last, this instrument can be used to orient single crystals based on its bulk structure, whereas X-ray Laue scattering is sensitive to the surface structure which may be different from the whole material one.

Similarly to the imaging station above, care must be given to the shielding of the experimental area, both by reducing the beam section at the beam port outlet and before the sample, installing a sapphire filter [19] in between the slits, as well as surrounding the instrument with heavy concrete blocks.

We have simulated a Laue diffractometer for single crystal sample installed on a tangential neutron beam port. The beam port section is diaphragmed to $ϕ 1 cm$ by mean of a small slit. A second slit is located at $L_{2} = 1 m$ from the first one, just before the sample location. A 10 cm thick sapphire filter is installed between the pair of slits. The simulated neutron flux at the sample location is found to be 10⁷ n/cm²/s for a 750 kW reactor power, which is sufficient to record Laue patterns within typically an hour. The detector is made of CCD panels arranged in an octagon, with a compact geometry at $L_{3} = 30 cm$ from the sample. The instrument geometry is shown in Fig. 11 and main parameters are presented in Table 3. The model makes use of the single crystal McStas sample component [31].

Fig. 11.

Overview of the Laue diffractometer model generated by McStas. The neutron beam comes from the left side through the collimator (length 3 m), and is filtered by a slit and a sapphire single crystal. The beam is reduced to the size of the sample immediately before it, and is then scattered by the sample. The neutrons are collected by CCD panels positioned in an octagon compact geometry around the sample. The heavy concrete shielding is represented as a 2 m block.

Table 3

Neutron Laue diffractometer parameters. The reactor power is set to 750 kW

Neutron Laue diffractometer
Beam port	Tangential, full opening
Slit1 at beam port	$ϕ D = 1 {cm}^{2}$ at $L_{1} = 3 m$ from core
Single crystal sapphire filter at beam port	Yes, 10 cm thick
Slit1-Slit2/sample distance	$L_{2} = 1.0 m$
Sample-detector distance	$L_{3} = 0.3 m$
Flux at sample [n/cm²/s]	10⁷
Detector, ZnS(Ag)–⁶LiF	7 panels, 20 × 20 cm²
Intensity at detector [n/s]	2 × 10⁷ (ideal)

The Laue pattern simulated using a large leucine protein (2-Amino-4-methylpentanoic acid, C₆H₁₃NO₂) single crystal sample is shown in Fig. 12. Bragg spots are large, due to the large slit and sample size. However, spots are well separated, allowing to perform a structure refinement. Similar patterns can be obtained with inorganic crystals, with however generally fewer spots.

Fig. 12.

Simulated Laue pattern from a leucine protein single crystal obtained on an ideal cylindrical detector (top), and on 7 lower sensitivity ZnS(Ag)–⁶LiF CCD panels (bottom), for a one hour exposure. The incoming beam is filtered with a sapphire single crystal. Intensity levels range from black to white.

The collected data can further be analysed using the FullProf [22] and Esmeralda [23] software.

7. SANS

The area available for the installation of neutron scattering instruments around a low power research reactor is often limited. A proper small angle neutron scattering instrument, which is typically 10 m long, is then hardly installable at such a facility. However, a small momentum transfer diffractometer can be envisaged. The measured signal is, again, derived from the Bragg law $\begin{matrix} 2 d sin θ = n λ, \end{matrix}$ where λ is the incident neutron wavelength, n is the order of the scattering, $2 θ$ is the scattering angle at the sample and d is the distance between the scattering unit planes. However, the instrument geometry assumes here that the scattering angle $2 θ$ remains limited, typically up to 5–6°, and a monochromatic incoming neutron wavelength is used, just as for the powder diffractometer above. Then the scattering condition requires the characteristic distance d between the scattering units to be large, e.g. 50–150 Å, which scatter as Debye–Scherrer cones for isotropic density systems similarly to powders. Such an instrument is suited to study colloidal suspensions, such as detergents, polymers with short chains, clays, and gels.

In order to estimate the potential characteristics of such an instrument, we have simulated a compact small angle scattering instrument operating at $λ = 4.74 Å$ from a highly oriented pyrolytic graphite PG(002) monochromator. The monochromatic beam has a 89.9° take-off angle which is convenient for building the concrete shielding around the neutron optics. The sample location is located about a meter from the monochromator, and the beam divergence, which is the key parameter defining the instrument resolution, is defined by two $ϕ 1 cm$ slits. In between the slits, a 10 × 2 × 2 cm polycrystalline Berylium filter, potentially cooled, can be advantageously installed to get rid of the thermal and fast neutrons [10]. Due to the narrow beam, the neutrons reaching the sample position originate from a small area on the monochromator surface, typically 2 × 2 cm², so that even if the monochromator is curved when shared with the powder diffractometer, the small angle instrument only uses the central slab of the device, and is not affected by the curvature. Alternatively, a small dedicated monochromator on a goniometer can be installed. The sample should be installed on a translation stage for easy alignment. The detector is modelled as 4 CCD panels located at $L_{3} = 2 m$ away after the sample with a nominal efficiency of 9.2%, assuming ZnS(Ag)–⁶LiF scintillator membranes. The $L_{3}$ distance can also be reduced down to $L_{3} = 0.3 m$ to increase the momentum range and start to record the first diffraction rings. The instrument is shown in Fig. 13, and main parameters are listed in Table 4. The McStas model uses the hard sphere colloidal suspension sample component [31]. The Be filter is simulated with the generic filter or the powder component. The flux at the sample position is found to reach 10⁴ n/c/cm² for a 750 kW reactor power.

Fig. 13.

Overview of the neutron compact small angle scattering instrument model generated by McStas. The neutron beam comes from the left through the collimator (length 3 m). It may optionally be filtered by a sapphire filter at the beam port outlet. The beam is reflected at $λ = 4.74 Å$ by a small PG(002) monochromator. The beam is further collimated with two $ϕ 1 cm$ cm slits separated by $L_{2} = 1 m$ , and is then scattered by the sample. The scattered neutrons are collected by 4 CCD panels arranged in a plane. The main parameters are detailed in Table 4.

Table 4

Compact small angle scattering instrument parameters. The reactor power is set to 750 kW

Compact small angle neutron scattering instrument
Beam port	Radial, full opening
Beam port diaphgram [width × height]	2 × 2 cm² (optional if shared with diffractometer)
Single crystal sapphire filter at beam port	Yes, 10 cm thick
Monochromator size [width × height]	2 × 2 cm²
Core-Monochromator distance	$L_{1} = 3.3 m$
Monochromator mosaicity	$η = 50 arcmin PG (002)$
Take-off [°]	$2 θ = 89.89 °$ ( $λ = 4.74 Å$ )
Monochromator-sample distance	$L_{2} = 1.0 m$
Monochromator-sample collimation	Two slits, $ϕ 1 cm$
Be powder filter	Yes, 10 cm thickness
Sample-detector distance	$L_{3} = 2 m$
Flux at sample [n/cm²/s]	10⁴
Detector, ZnS(Ag)–⁶LiF	4 panels, 20 × 20 cm²
Intensity at detector [n/s]	1 (ideal)

The instrument parameters listed in Table 4 correspond to a compact configuration, which provides a flux at the sample in agreement with what can be obtained at the existing SANS in Malaysia [28]. The flux at the detector is expected to be low, with measurement times typically of a few hours.

We present the simulated scattering pattern recorded by the detector in Fig. 14, from a $R = 50 Å$ hard sphere colloidal suspension. The radial integration can then be converted to the intensity vs. momentum, $I (q)$ .

Fig. 14.

The scattering pattern from a 50 Å hard sphere colloidal suspension simulated on the 4 CCD panel detector with the compact small angle scattering instrument operating at $λ = 4.74 Å$ ( ${log}_{10}$ scale, dark to light grey intensity). The radial integrated signal is shown on the right side, and can further be treated and analysed to derive the microscopic information about the nature and shape of the scattering units.

A cost effective installation can be obtained by positioning the SANS small monochromator after the powder diffractometer one, thus sharing the same sapphire filter, and concrete shielding. Also, the powder diffractometer operating at $λ = 2.36 Å$ , which is about half of what is needed by the SANS at $λ = 4.74 Å$ , will extract neutrons with wavelength $2.36 / n$ where n is an integer, and thus act as a filter for the SANS. The beryllium filter would then be optional. This arrangement is shown in Fig. 15.

Fig. 15.

Overview of the shared neutron beam port used to installed the large powder diffractometer curved monochromator, followed by the small SANS flat monochromator. In this arrangement, the powder diffractometer acts as a higher order filter for the SANS, improving the signal to noise ratio. The axes ticks are in meters.

In a further evolution, a low pressure ³He gas PSD detector should be preferred to minimize the sensitivity to the background originating from the reactor. The collimation length and the detector should also advantageously be installed in evacuated volumes to reduce scattering from air.

The recommended software to treat the data is LAMP [21] and GRASP [7], whereas the structure and form factors can be analysed with e.g. SASView [25] or SASfit [5].

8. Instrument parts

As detailed in the instrument geometries above, a number of neutron optics parts must be purchased for the installation of the instruments. However, the detectors, the filters and the monochromators can be shared between the different instruments, allowing a cost effective solution, which also ensures a better maintenance with exchangeable parts.

Filter Be: 10 cm thickness 10 × 3 × 3 cm³ for the SANS, in between the slits after the monochromator and before the sample. Powder is packed into a sealed Al box. Cooling is optional (bringing 20% more transmission). Must be surrounded by B₄C top/bottom/left/right. Provider: Sigma Aldrich, 10$/g.

Filter HOPG: 5 cm thickness 5 × 4 × 4 cm³ for the powder diffractometer, between the monochromator and the sample. Must be surrounded by B₄C top/bottom/left/right. The filter requires to be stacked and perfectly oriented (within e.g. 0.1°). Such a filter should be ordered from other facilities or specialised companies.

Filter Sapphire: 10 cm thickness made of few stacked single crystal pieces. To be installed at most neutron beam ports to scatter out fast neutrons. Can be installed directly inside the neutron beam port outlet to benefit from the reactor concrete shielding. Must be surrounded by B₄C top/bottom/left/right. Orientation is not detrimental. Provider: Stettler Sapphire and GT Advanced technologies.

B₄C rubber: made in-house by pouring B₄C powder into natural or silicon rubber in a 20–40% wt. fraction [12]. Thickness 5 mm. Can be used to surround optics, glued over concrete shielding blocks and mechanical parts. Provider: Sigma Aldrich, 1$/g.

Structural parts: better made of Aluminium (AG3 NET, 5754 alloy) which has a low neutron activation.

Monochromators: needed for the large 13 × 13 cm² diffractometer curved monochromator, and the small 2 × 2 cm² SANS flat monochromator. These optics must be mounted on remote-controlled translation/rotation stages. Provider: Optigraph or Momentive or Agar Scientific or Panasonic, 100–200$/cm² for 2 mm thickness.

Slits: Aluminium plate with B₄C rubber on both sides, and a hole in the centre.

Mechanics: Rotation and translation stages, motorised, from Edmund Scientific or Newport (about 5 k$/stage including remote controller). These stages are needed for accurate positioning of the monochromators, at well as the orientation of the sample at the Laue diffraction and imaging instruments. Such stages can also be used for shutters.

Detector: ZnS(Ag)–⁶LiF scintillators from e.g. Neutron Optics or Photonic Science (3 k$/400 cm²). These detectors are portable between the experiments, and easy to use. They are suited where a small pixel size is needed, e.g. at the Laue diffraction and imaging instruments. An alternative, especially for the diffractometer and the SANS requiring low background, is ³He PSD proportional counter panels with low pressure from e.g GE Reuter Stokes (50–100 k$/m²).

Beam stop: to be located after the sample, made from B₄C powder in an Al casing. It should be moved further down-stream to minimise the gamma radiation originating from the neutron absorption by boron.

Beam shutters: motorised translation stages to be installed at the neutron beam port outlets, holding a 14 cm wide, 10 cm thick B₄C rubber/lead absorber to close the beam.

9. Shielding

The structural shielding around a research reactor is usually made of concrete, enriched with iron. At neutron scattering instruments, it is desirable to minimised any radiation source except that being measured, essentially to protect the scientists using the instruments and improve the signal to noise ratio.

The background signal includes:

fast neutrons from the reactor;

gamma radiation from both the reactor, but also from absorbed neutrons at slits, shutters, walls;

scattered neutron from surrounding instruments and walls.

The shielding around experiments should at the same time absorb gamma radiation, by involving heavy elements, and absorb neutrons by capture in high neutron absorption cross section elements. However, as the neutron absorption cross section varies as $1 / \sqrt{E}$ outside of absorption resonances, where E is the neutron energy, it is desirable to thermalise neutrons before their capture, i.e. slow them by removing kinetic energy. The thermalisation can efficiently be performed by materials with high scattering cross sections, such as hydrogen, contained in organic molecules and e.g. water. Consequently, an efficient shielding assembly can be made from paraffin and concrete, with an additional lead layer where appropriate.

We propose the following composition for the shielding materials:

Baryte-colemanite heavy concrete can be manufactured in-house from 79% wt. baryte BaSO₄ (contains Ba, gamma absorber), 7% wt. Portland cement., 10% wt. colemanite CaB₃O₄(OH)₃ · H₂O (contains boron, neutron absorber), 4% wt. water [11,18]. Walls should surround all experimental areas with a chicane for access, as e.g. 20 cm thick layer on the sides, and 40 cm down-stream direct beam.

Borated paraffin, adding 5–20% wt. boric acid B(OH)₃, or 5% wt. B₄C powder can be manufactured and melted in-house into the desired shape, with thickness e.g. 5 cm at low-medium power research reactors. The wax plates can be housed within two thin Al plates for easier handling. To be used around detectors and screwed onto both sides of concrete walls.

Lead blocks, 5 cm thickness. To be mostly used around the neutron beam port outlet, in shutters, and down-stream direct beam. As lead emits photo-neutrons, this layer should better be squeezed in between concrete and borated paraffin.

All shielding blocks should be shaped in order to suppress the direct view between blocks, with an interlock geometry. V shaped lead blocks are common, and allow to raise wall protections. Concrete blocks can be V or T shaped, for easy assembly.

10. Conclusion

We have demonstrated that simple, cost effective neutron scattering instruments can be installed at low-power research reactors. These instruments are a powder diffractometer, a small angle scattering instrument, a Laue camera and a radiography station, which can be optionally grouped to share beam ports and neutron optics parts. This way, the local neutron scientist community can make use of these facilities to study geological samples, metallurgy alloys, ceramics, archaeology artefacts and industrial imaging, small organic crystals, colloidal suspensions and polymers. In addition, these instruments can be used for teaching purposes, to develop the neutron scattering methodology. We estimate that most of these instruments can be installed for a budget lower than 200 k$. In practice, the flux estimates obtained in the scope of this study should be considered as higher values, as the model instruments are assumed to be perfect. We thus estimate that, once built, the instrument flux at the sample positions should be lowered by e.g. 30%, as often seen in similar studies [9,16,32,33].

Instrument models are included as metadata for use with McStas release 2.2, and for a 750 kW thermal reactor source.

References

I.A.

Alnour,

Wagiran,

Ibrahim,

Hamzah,

B.S.

Wee,

M.S.

Elias and

J.A.

Karim, J. Radioanal. Nucl. Chem. 296 (2013), 1231. doi:10.1007/s10967-012-2375-9.

Bergsma and

Van Dijk, Nucl. Inst. Meth. 51 (1967), 121. doi:10.1016/0029-554X(67)90371-0.

Böck,

Hasegawa,

Jericha,

Steinhauser and

Villa, Journal of Energy and Power Engineering 7 (2013), 654.

Borio di Tigliole,

Cammi,

Chiesa,

Clemenza,

Manera,

Nastasi,

Pattavina,

Ponciroli,

Pozzi,

Prata,

Previtali,

Salvini and

Sisti, Prog. Nuc. Energy 70 (2014), 249. doi:10.1016/j.pnucene.2013.10.001.

Bressler,

Kohlbrecher and

A.F.

Thunemann, J. Appl. Crystallogr. 48 (2015), 1587. doi:10.1107/S1600576715016544.

R.-C.

Chen,

Dreossi,

Mancini,

Menk,

Rigon,

T.-Q.

Xiao and

Longo, J. Synchrotron Rad. 19 (2012), 836. doi:10.1107/S0909049512029731.

Dewhurst, http://www.ill.eu/lss/grasp/.

Farhi,

Debab and

Willendrup, J. Neut. Res. 17 (2013), 5.

Farhi,

Monzat,

Arnerin,

Van Vuure,

Castán-Guerrero,

Hennane,

P.A.

Harraud,

Campioni,

Fuard,

Ollivier and

Willendrup, Journal of Neutron Research 17 (2013), 63.

10.

Freund, Nucl. Inst. Meth. 213 (1983), 495. doi:10.1016/0167-5087(83)90447-7.

11.

R.B.

Gallaher and

A.S.

Kitzes, ORNL report 1414, 1953.

12.

S.E.

Gwaily,

M.M.

Badawy,

H.H.

Hassan and

Madani, Polym. Test. 21 (2002), 129. doi:10.1016/S0142-9418(01)00058-7.

13.

IAEA, https://nucleus.iaea.org/RRDB/RR/ReactorSearch.aspx.

14.

Kennedy and

R.C.

Eberhart, Particle swarm optimization, in: Proc. IEEE Conf. on Neural Networks, IV, Piscataway, NJ, 1995, pp. 1942–1948.

15.

A.C.

Larson and

R.B.

Von Dreele, Los Alamos National Laboratory Report LAUR 86-748, 1994.

16.

Lefmann and

Nielsen, Neutron News 10 (1999), 20. doi:10.1080/10448639908233684.

17.

Mastronarde,

Kremer,

Gaudette,

Held,

Xiong and

Heumann, Univ. Colorado, http://bio3d.colorado.edu/imod/.

18.

Mesbahi,

Alizadeh,

Seyed-Oskoee and

A.-A.

Azarpeyvand, Annals of Nuclear Energy 51 (2013), 107. doi:10.1016/j.anucene.2012.07.039.

19.

D.F.R.

Mildner,

Arif,

C.A.

Stone and

R.K.

Crawford, J. Appl. Cryst. 26 (1993), 438. doi:10.1107/S0021889893000433.

20.

NiftyRec, http://niftyrec.scienceontheweb.net/wordpress/.

21.

Richard,

Ferrand and

G.J.

Kearley, J. Neut. Research 4 (1996), 33. doi:10.1080/10238169608200065.

22.

Rodriguez-Carvajal, Physica B 192 (1993), 55. doi:10.1016/0921-4526(93)90108-I.

23.

Rodriguez-Carvajal and

Fuentes-Montero, http://lauesuite.com/.

24.

J.R.

Santisteban,

Edwards,

M.E.

Fizpatrick,

Steuwer and

P.J.

Withers, Appl. Phys. A 74 (2002), S1433. doi:10.1007/s003390101241.

25.

SASView, http://www.sasview.org/.

26.

Snoj and

Smodiš, Nuclear Energy for New Europe (2011).

27.

SPIERS, http://spiers-software.org/index.htm.

28.

M.A.M.

Sufi,

Radiman,

Wiedenmann and

Mortensen, J. Appl. Cryst. 30 (1997), 884. doi:10.1107/S0021889897001738.

29.

B.H.

Toby, J. Appl. Cryst. 34 (2001), 210. doi:10.1107/S0021889801002242.

30.

H.R.

Vega-Carrillo,

V.M.

Hernández-Dávila,

Aguilar,

Paredes and

Rivera, Applied Radiation and Isotopes 83 (2014), 252. doi:10.1016/j.apradiso.2013.04.008.

31.

Willendrup,

Farhi,

Knudsen,

Filges and

Lefmann, Component Manual for McStas, 2014. ISBN 978-87-550-3680-2.

32.

Willendrup,

Farhi and

Lefmann, Physica B 350 (2004), 735. doi:10.1016/j.physb.2004.03.193.

33.

Willendrup,

Filges,

Keller,

Farhi and

Lefmann, Physica B Cond. Matt. 385 (2006), 1032. doi:10.1016/j.physb.2006.05.329.