Advanced Small-Angle Scattering Instrument Available in the Tokyo Area. Time-Of-Flight, Small-Angle Neutron Scattering Developed on the iMATERIA Di ﬀ ractometer at the High Intensity Pulsed Neutron Source J-PARC

: A method of time-of-ﬂight, small-angle neutron scattering (TOF-SANS) has been developed based on the iMATERIA powder di ﬀ ractometer at BL20, of the Materials and Life Sciences Facility (MLF) at the high-intensity proton accelerator (J-PARC). A large-area detector for SANS, which is composed of triple-layered 3 He tube detectors, has a hole at its center in order to release a direct beam behind and to detect ultra-small-angle scattering. As a result, the pulsed-neutron TOF method enables us to perform multiscale observations covering 0.003 < q (Å − 1 ) < 40 ( q max / q mix = 1.3 × 10 4 ) and to determine the static structure factor S ( q ) and / or form factor P ( q ) under real-time and in-situ conditions. Our challenge, using unique sample accessories of a super-conducting magnet and polarized neutron, is dynamic nuclear polarization (DNP) for contrast variation, especially for industrial use. To reinforce conventional SANS measurements with powder materials, grazing-incidence small-angle neutron scattering (GISANS) or reﬂectivity is also available on the iMATERIA instrument.


Introduction
The softness of materials such as gels, foams and rubber is attributed to the reorganization of self-assembled molecular structures [1]. The interaction energy to associate molecules (hydrogen bonding, van der Waals or hydrophobic interactions) is of the order of the thermal energy at room temperature (~k B T). Furthermore, molecules comprising soft materials possess a large amount of internal degrees of freedom. Consequently, soft materials react in a viscoelastic manner to external stimuli such as deformation (stretching or shear), and/or temperature or pressure jumps. The taste and texture of foods, the softness of clothes, and the fuel efficiency and stopping performance of tires, are all related to viscoelasticity coupled with the hierarchical structure. Time-of-flight (TOF) neutron SANS instruments across the world have been developed considering the type of neutron sources (reactor or accelerator) [3]. Research reactors with a continuous neutron beam are advantageous for large-scale observation by achieving the smallest accessible wave number q min . A focusing ultra-small-angle, neutron-scattering method was recently developed by using a monochromatic beam and focusing lenses [3,4]. The refractive lens [5,6] or magnetic lens [7] was invented in around 2000 and is combined with a small area detector with a higher spatial resolution of mm. The focusing method covers the q-region down to 10 −4 Å −1 and reinforces a Bonse-Hart double-crystal method using perfect crystals [8,9], which covers the q-region down to 10 −5 Å −1 [3,4]. For example, at the research reactor JRR-3, the focusing SANS instrument (SANS-J-II) [3,4] and double-crystal, ultra-small-angle scattering (USANS) instrument (PNO) [10] are combined complementarily. The focusing method is standard as an option on pin-hole SANS instruments all over the world [11][12][13][14][15][16]. Hierarchical structures, appearing in various types of soft matter, have been explored by combining SANS and USANS instruments at the JRR-3 reactor, Japan [17][18][19][20].
A proton accelerator, providing pulsed neutrons, is advantageous if one wishes to extend the q-region wider, achieving an accessible maximum wave number q max . A time-of-flight (TOF) method does not require monochromation and minimizes the loss of incident neutrons [21,22]. A wide range of length scales is covered from atomic to meso scales. Even at reactors, the incident beam can be chopped as a pulse, in order to achieve q max /q mix = 1000 [23,24]. Time-of-flight or wavelength dispersion has the merit of avoiding angular scanning, for changing q of reflectivity or double crystal USANS [25,26]. Recently compact pulse sources have become available. The ib-SAS instrument, using a TOF method with pulsed neutrons, was prepared as a RANS compact neutron source at RIKEN, Wako Japan [27,28].
In this paper, we report on recent developments and the current status of time-of-flight SANS on the iMATERIA diffractometer at the Japan Proton Accelerator Research Complex (J-PARC). The available wavelength is ranged from 1 to 10 Å, when we choose double-frame chopper operation. Consequently, we achieved multiscale observations covering 0.003 < q (Å −1 ) < 40 (q max /q mix = 1.3 × 10 4 ). A new area detector for SANS was installed, in order to provide access for USANS in the future [29]. Reflectivity is also available to assess fine structures on a solid surface [30,31]. Our big challenge at BL20 is to achieve contrast variation by dynamic nuclear polarization (DNP); both the neutron beam and hydrogen (in the specimen) are polarized in order to tune the coherent scattering length density for contrast variation. The DNP apparatus at the iMATERIA diffractometer is already open for industrial use [32]. There are a variety of SANS instruments achieved currently, which are listed in the introduction . In this paper, we intend to indicate a new benchmark of advanced SANS apparatus for the next generation.

iMATERIA Spectrometer
At the Japan Proton Accelerator Research Complex (J-PARC), three MeV protons hit a mercury target with a pulse repetition frequency of 25 Hz. In the Materials and Life Sciences Facility, (MLF), there are 23 beam lines in two beam halls, and the iMATERIA diffractometer is located at BL20 (Figure 2), facing towards a poisoned decoupled moderator producing a narrow pulse width of~10 µs. Neutrons are transported by a supermirror (3 θ c ) guide. Along the beam line, there are three disk choppers (at 7.5, 11.25 and 18.75 m) and a T 0 chopper (10 m) in order to shape the neutron pulse in time. The supermirror guide after the disk chopper 2 is 11.75 m long and 34 × 34 mm 2 in cross-section. Upstream of the instrument (part A in Figure 2), a polarizing supermirror, π-flipper and multipinhole collimator have been installed [32], to make finer collimation possible for USANS. At the downstream of the instrument (part B in Figure 2), we will install a USANS detector in the future.  Figure 3a shows a schematic diagram of the iMATERIA diffractometer (BL20), which is dedicated to industrial use. The iMATERIA diffractometer was originally designed for powder diffraction, and its original design before reconstruction for SANS is described in [33]. Four detector blocks-(1) small-angle (SA; 0.5 < 2 θ < 5.5 • ), (2) low-angle (LA; 10 < 2 θ < 40 • ), (3) 90 degree (SE; 80 < 2 θ < 100 • ) and (4) backscattering (BS; 150 < 2 θ <175 • )-surround the sample position and cover a wide 2 θ range (0.1 • -180 • ). Figure 3b shows the 3-dimensional layout of the 3 He detectors, the total number of which is about 1500. The view from sample position downstream is towards the small-angle detectors in the center and a six-fold array of low-angle detectors covering the forward-scattering path. Since 2016, the large-area detector for SANS has been refurbished [28] and the instrument started its User program. The details of the large-area detector are described in [29]. Figure 4a shows a photo of the large area detector, composed of 3 He linear position-sensitive detectors. With a three-layered structure (Figure 4b), we prepared a hole at the center (100 × 100 mm 2 ), in order to release the direct (unscattered) neutron beam. Behind the small-angle detector as shown by B in Figure 2, we installed an additional chamber containing the 3 He direct beam monitor and USANS detectors (Figure 4c). With the USANS apparatus, we aim to access up to 10 −4 Å −1 , in the near future. As a result, the versatile combination of SANS and powder diffraction is characteristic for wide q observation with high resolutions. By combining wide range of scattering angle (2 θ) and a wide band of wave length (λ = 1-10 Å), (combination of scattering-angle-dispersive and wavelength -dispersive methods), a wide q-region (from 0.003 to 40 Å −1 , implying q max /q mix = 1.3 × 10 4 ) can be immediately covered. (c) Ultra-small-angle scattering detector is developed behind the main small-angle scattering detector, as indicated by B (corresponding to B in Figure 2). Figure 5 shows a schematic diagram showing scattering geometry (incident and scattered neutron with wave number 2 π/λ and scattering angle 2 θ). The orthogonal coordinate is defined with the z-axis parallel to the incident beam, and x-and y-axes parallel to the surface of the small-angle detector. A scattering process is described by scattering vector → q . For glassy materials or a solution with slow dynamics, we postulate elastic scattering wave number k for the incident and scattered beam, Therefore, we obtain

Time-of-Flight Scattering Method
where λ is a wavelength and 2 θ is the scattering angle. For incident and scattered beams, in orthogonal coordinates, where L 2 is a distance from sample to detector, as defined in Figure 3b. We redefine or with polar coordinates where φ is an azimuthal angle, defined in the x-y plane. The neutron is a fundamental particle with mass (m n = 1.675 × 10 −27 kg) and spin (P n = ±1/2). According to a textbook of quantum mechanics and wave-particle duality therein, the neutron wavelength is given by the de Broglie principle with Plank's constant h, where p is momentum given with mass and velocity v, as p = m n v. With a traveling distance and time, L and t TOF , velocity is given as v = L t TOF . Therefore, we obtain For scanning q, there are two methods: (i) angular-dispersive method-by changing a scattering angle 2 θ (usually detector positions with respect to the beam), q values are scanned. For SANS at a reactor source, we usually change detector positions away from the sample position with using continuous and monochromatic neutrons, and (ii) wavelength-dispersive method-for SANS at a spallation source, q values are usually scanned by using the time-of-flight method using highly intensified nonmonochromatic neutrons. Note that even at a spallation source with the pulsed neutrons or at a reactor source [23,24], the combination of (i) and (ii) is more effective to widely scan q. To elucidate hierarchical structure, many orders of magnitude in q should be covered as much as possible. Toward wider q, the wavelength-dispersive method at a pulsed neutron source profits by gaining q max and therefore q max /q min . Toward smaller q, on the other hand, the wavelength-dispersive method is no longer necessary because the smallest q is obtained by a combination of the smallest θ and longest λ. Therefore, pin-hole SANS (not a perfect crystal type) is preferable for achieving the smallest q min by avoiding frame-overlap of pulsed neutrons. Figure 6 shows a data reduction program of SANS at the iMATERIA spectrometer, treating time-of-flight event data, which converts the real space image to a reciprocal space image by considering Equations (1)-(6), (parts 4 and 5 in Figure 6). By selecting a narrow band of wavelengths, users can examine the scattering as a monochromatic beam experiment like at reactor (part 2 in Figure 6). According to Equations (1)-(6), from event data with a flight time t TOF , q-value is determined. By circularly averaging or sector averaging over an azimuthal angle ϕ (part 3 in Figure 6), the scattering profile I s obs (λ, θ) can be obtained. The data-reduction program can also be applied to reflectivity measurements, which will be discussed in Section 5.2.3. The measured intensity I s obs (λ, θ) is given with background scattering I bck obs (λ, θ) as follows:

Data Reduction for Plate Shape Sample
where dΣ dΩ (q) is the differential scattering cross-section. Note that I s obs (λ, θ) is proportional to the incident beam intensity I 0 (λ), sample volume (A S D S ) and transmission T r s (λ), where A S and D S is sample cross-section and thickness. K I is an instrument constant. η(λ) and ∆Ω are counting efficiency and the solid angle of a detector element. Therefore, we obtain Note that I 0 (λ), T r s (λ) and η(λ) are wavelength-dependent. We define a distance as L = L 1 + L 2 , which is a sum of L 1 (from T 0 chopper to a sample) and L 2 (from a sample to detector). An absolute intensity unit (cm −1 ) is obtained by considering SANS for a secondary standard of glassy carbon [29]. With the neutron counts from a direct beam monitor behind a small-angle scattering detector, neutron transmission T r s (λ) is calculated as follows: where I s 1 (λ), I bck 1 (λ) and I close 1 (λ) are the transmitted beam intensity for the sample, background and beam-close measurements, respectively. The beam-close measurement is conducted by putting B 4 C block at the sample position. D s is thickness of the sample. Σ T is macroscopic total cross-section, composed of scattering Σ S , incoherent scattering cross-section (Σ inc ) and absorption cross-section Σ A , as follows: Using coherent and incoherent scattering and absorption microscopic cross-sections (σ coh , σ inc , and σ abs , respectively), Σ T is given by Σ T = n(σ coh + σ inc + σ abs ), where n is number density as defined n=N/V. Microscopic scattering cross-sections are given by σ coh = 4πb 2 coh and σ inc = 4πb 2 inc , where b coh and b inc are coherent and incoherent scattering lengths. Note that a mean free path (MFP) is given by Differential scattering cross-section dΣ dΩ (q) is given by is a vector between i-th and j-th atoms. Phase difference in path between the neutrons scattered by i-th and j-th atoms is given as Normal SANS usually observes an isotropic and random-oriented powder or solution. In this limit, the golden equation by Debye [34], is obtained as follows: Then it becomes where Θ is an angle between a scattering vector and principle axis of the scattering object. After random-averaging in Θ for isotropic systems, q is given as scalar. Because of random-averaging in Θ, limited special resolution accessible by SANS does not allow the study of local ordered structures, for example, contact ionomers on a fuel cell catalyst. To address this issue, the option of simultaneous use of reflectivity (RF) technique is available on the iMATERIA (see Section 5.2.3 for more details). Our data reduction program in Figure 6 is also available for obtaining RF profiles. It should be strongly emphasized that our measurable SANS is determined simply by varying q and b, according to Equation (13), in relation to materials we are interested in.

Sample Environment
The sample position of the iMATERIA diffractometer is in vacuum and 0.6 m below the floor of the standing stage for users. In order to handle a sample specimen for SANS measurements, we prepared an elevator stage with a 4-fold slit with sintered B 4 C plates with a tapered edge (Figure 7a,b). Figure 7c is a top view of the stage: it is sustained by four poles which do not block the scattering toward four detector blocks from small-angle to backward scattering. On the stage, we can change between different sample environments, e.g., a temperature-controlled sample stage (25~200 • C) with five sample positions. Figure 7d shows the sample stage kept in air. Figure 8 shows the auto-sample changer for round-shaped specimens operated at room temperature (no temperature control); 40 cuvettes automatically are moved to the beam position. The elevator stage in an air chamber is also available, which is used for in-situ grazing-incidence scattering (GISANS) observation of polymer electrolyte fuel cells under operation, as discussed below in Sections 5.2.3 and 5.2.4. Figure 9 shows the gonio-stage for GISANS and reflectivity available on a sample stage ((b) horizontal and (c) vertical goniometer stages, respectively). It is shielded in an atmosphere environment container (Figure 9a). Temperature is controlled by heaters. To control the substrate angle relative to the incident beam, a rotary goniometer stage is used. As an extreme condition measurement, dynamic nuclear polarization (DNP) with a super-conducting magnet, with 4 He cooler, is also available on iMATERIA (see Section 5.2.5).    Figure 10a shows incoherent scattering from a vanadium standard specimen, calibrated for counting efficiency for different wavelengths. The different colors indicate the incoherent scattering obtained with different wavelengths λ from 1 to 10 Å. Figure 10b shows I 0 (λ), obtained for double-frame (12.5 Hz) operation. After calibration of I 0 (λ), the scattering obtained for individual wavelengths λ is normalized in order to have q-independent incoherent scattering, as shown in Figure 10a. The determined calibration factor η(λ) is shown in Figure 10c,

Incoherent Scattering of Vanadium
which is used for data reduction according to Equations (7)-(9).  Figure 11a shows the result obtained for a powder of mesoporous silica (TMPS 2.7 mm), showing the q-region covered by a single shot measurement simultaneously using the forward detectors (small-angle, low angle detector and partially-ultra-small-angle scattering blocks) (0.003-4 Å −1 ). It should be stressed that the minimum observable wave number q min = 0.003 A −1 . When we access down to q min = 0.003 Å −1 , we then use a USANS detector and insert fine collimation upstream of the sample (δθ = 0.004 rad for both vertical and horizontal directions). Figure 11b shows the wavelength-dependent transmissions ( T r s (λ)) obtained for the mesoporous silica (TMPS-2.7 mm). According to Equations (9) and (10), T r s (λ), which varies from 1.0 to 0.8 as increasing λ, is given by incident beam intensity (I s 0 (λ) and I 0 (λ)). Absorption cross-section depends on wavelength λ.

Mesoporous Silica
Note that σ abs is determined for thermal neutrons of λ = 1.8 Å. The gradual decrease in T r s (λ) is attributed to neutron absorption and diffraction.

Mono-Disperse Microspheres (SiO 2 )
In order to examine q-resolution, we measured SANS on SiO 2 microspheres, as shown in Figure 12. The scattering profile (red in Figure 12) was obtained by a single shot measurement simultaneously using the forward detectors (small-angle and low-angle detector blocks). The scattering function for a spherical object is given by a form factor P(q) and a structure factor S(q), dΣ dΩ (q) = ∆b 2 n p V 2 p P(q)S(q) where n p and V 2 p are number density and volume of a sphere. ∆b is a contrast as defined, ∆b = b j − b i . Note that in case of air, b i = 0. The difference in scattering length density, the form factor P(q) of sphere with a radius R is given by, When there is polydispersity in size R, the averaged scattering is given bŷ The size distribution is described by a Gaussian function with a standard deviation σ R . C is a normalization factor. L(q) is, for example, a Debye model with hard spheres (see [3]). The q-resolution is given by which is composed of an angular resolution term and a second wavelength resolution term. Scattering obtained for SiO 2 microsphere, for examining the q-resolution. I ex (q) (black symbol) is the scattering profile experimentally determined.P(q) is smeared by the size distribution of spheres given by Equation (23) and I ob (q) is smeared by both size distribution and instrument angular resolution. The broken line is a guide for the eye showing incoherent scattering.
In the low-q limit, the first term (cot θ∆θ) 2 is dominant, whereas at high q, the second term ∆λ/λ is dominant (in case of iMATERIA, ∆λ/λ = 0.00043). ∆θ is the beam divergence of incident beam, which is determined by the combination of collimator slit widths S 1 and S 2 and the detector element size.
The smeared scattering is given bŷ P(q) is an analytical scattering function for a spherical object, after averaging over a size distribution. The resolution function is given by a gaussian function [35], as follows: and In case of time-of-flight (wavelength-dispersive method), we sum up the scattering functions with different wave lengths λ max > λ > λ min which gives a certain range of q. Figure 12 shows the observed SANS I ob (q) and the analysis including the size distribution and q-resolution with δθ = 0.0025 rad and δλ/λ = 0.00043, which is common for a powder diffractometer at pulsed source. Note that I ex (q) is the scattering profile experimentally determined. The size distribution was determined to be R AV = 160 Å and σ R = 30 Å. In the region q > 0.04 Å −1 , I ob (q) is smeared by the size distribution. For q < 0.04 Å −1 , I ob (q) is affected by the angular resolution of δθ = 0.0025 rad, in addition to the size distribution. The discrepancy at low q may be due to inter-particle interference (L(q)) or to excluded volume effects. The deviation at high q is due to incoherent scattering, which is not included in the analytical model. The angular divergence of δθ = 0.0025 rad is determined by the supermirror tubes after the disk chopper; its length is 11.75 m and cross-section is 34 × 34 mm 2 gives δθ = 0.0028 rad.

Challenging Topics of Industrial Use
5.2.1. One-Shot Observation Covering a Wide Range of Length Scales (1 A~100 nm) Figure 13a shows a scattering curve obtained for the polymer electrolyte Nafion ® . Three sheets (each of thickness 50 µm) are packed in an aluminum holder, swollen by H 2 O or D 2 O, and dried. The small-angle detector covers a q-range of 0.007 < q <0.5 (Å −1 ), whereas the low-angle detector covers a q-range of 0.2< q < 5 (Å −1 ). The hierarchical self-assembled structure is recognized by the scattering maxima attributed to crystalline grains at q =0.055 Å −1 (~20 nm), water channels or ion cluster at q = 0.15-0.2 Å −1 (~nm) and crystalline chain order at q = 1.2 Å −1 . At higher q (>0.5 Å −1 ), incoherent scattering from hydrogen in the films appears in proportion to its concentration. The result was obtained at a proton beam power of 200 kW, with double-frame operation (80-ms time interval). The measurement time was 600 s When at J-PARC's full beam power of 1 MW, with single frame operation, we expect this to reduce to 60 s We will then be able to detect a hierarchical structure or transient states, during processing. After performing the measurements, the user can generate histogram event data on a personal computer (part 1 in Figure 6), in a flexible manner. Figure 13a,b demonstrate the results obtained by summing up different numbers of event data. Evaluation of an appropriate measurement time is as follows: for 8 s (100 flames), the profile exhibits large errors, whereas for 800 s (10,000 flames), the profile exhibits better statistics.

Time-Resolved and Wide Range of Length Scales: Vanishing Foam
In Figure 14, we report a time-resolved observation on a vanishing foam structure. A wide range of length scales (1 Å~100 nm) is covered by the single-shot measurement. The foam was prepared with an air pump, at 140 mm height from the liquid surface. Amino acid-type surfactant molecules (N-dodecanoyl-N-(2-hydroxyethyl)-β-alanine) were investigated (denoted as C 12 -βAla-C 2 OH). We prepared surfactant solution with D 2 O (100 mmol dm −3 ). Figure 14 shows time-resolved SANS obtained for C 12 -βAla-C 2 OH, after starting the pump (time interval is 1 min, with a 500 MW operation). From smaller q, we recognize small-angle scattering attributed to (1) single bimolecular layers with a disk shape, (2) micelles in a bimolecular layer and (3) incoherent scattering due to hydrogen atoms of surfactants. As the time proceeds, the SANS profile changes according to the arrows in Figure 2. We found the following: (i) Immediately after pumping a scattering maxima at around q = 0.075 Å −1 appears, attributed to the micelles in the foam. The interdomain distance (~80 Å) is identical to that found in the surfactant solution before pumping. (ii) In the lower q-region, at 300~360 s after pumping, crossover q-behavior from q −2 to q −4 appears at q c = 0.023 Å −1 . From q c , we estimated the thickness of foam as d = 270 Å (=2π/q c ). (iii) At higher q, incoherent scattering appears and its intensity decreases as time proceeds. This is due to water drainage and its dropping under gravity. The details of the date analysis are described in [36].

Fine Structure of Fuel Cell Catalyst, in Combination with Neutron Reflectivity
Normal SANS is performed with the incident beam along the z-axis, and we obtain → q = q x , q y , 0 . By circularly averaging in q, carried out with the position of the direct neutrons as the origin, a SANS intensity I(q) is acquired as a function of q = q 2 x + q 2 y . The powder of catalyst is packed in a plate-shape cuvette, i.e., an isotropic and spherically symmetric system (see Figure 15a). Random average on orientation in Θ gives the equation, the so-called Debye formalism in Equation (13). By using contrast variation via water exchange with H 2 O/D 2 O, we observed SANS from the catalyst powder [37]. As a result of a model analysis using a core-shell structure, we determined an ionomer thickness of about 5 nm and that the water content is 17 wt.%. It should be denoted that the intra-layer structure of the Nafion ® shell cannot be accessed by normal SANS. This is because of the random average on orientation in Θ and the Debye formalism.
To observe the local structure more clearly in the vicinity of the catalyst surface, we combine reflectivity measurements in order to reinforce the normal SANS. Reflectivity (specular reflection) is sensitive to the density profile (db/dz) along a direction of normal vector. Reflectivity is described by To model the surface structure, Pt is sputtered onto a flat silicon substrate (111) and the Nafion ® ionomer is spin-coated (see Figure 15a). Reflectivity was detected on iMATERIA using the reflection stages shown in Figure 9. As for the model catalysts prepared on Si substrate and humidified Nafion ® , we determined layer thickness, water-thin layer on Pt surface, water distribution in a Nafion ® thin layer and drying at surface. The details are given in [37].

Time-Resolved Total Reflection Measurement (Kinetic Measurement)
The time-of-flight wavelength-dispersive method enables us to perform time-resolved reflectivity within a short time interval, without having to move or rotate the sample. In order to determine mass transfer along the film thickness direction, we performed time-resolved total reflection measurements [31]. Figure 16a shows a schematic diagram of total reflection at the Fresnel surface between the bulk Nafion ® film and Si substrate, as reported in [31]. After exchanging water from H 2 O to D 2 O, we can detect the change in total reflection angle at the Fresnel surface behind the film. Figure 16b shows the time-resolved total reflection for water exchange from D 2 O to H 2 O with a 5-min time interval. Obviously, the critical wave number for total reflection (q c ) shifts from high to low q, and the q-behaviors in reflectivity curves becomes more rounded, which is attributed to the imaginary part of scattering length (incoherent scattering from hydrogen) at the Fresnel interface. After analyses of the time-resolved reflectivity, we evaluated the diffusion coefficient of water in a Nafion ® film swollen by water.

Spin Contrast Variation for Industrial Use
As mentioned in the introduction, DNP at the iMATERIA diffractometer is already opened for industrial uses: in [32], we report the first result obtained on the tire rubbers. For contrast variation, a deuteration method is normally used. One exchanges those hydrogen atoms bound to the molecule by covalent bonds, or one substitutes the solvent with a deuterated one. This method requires us to prepare a model system, simulating industrial products. It means that significant user effort and compromise are necessary. However, it is impossible to perform deuteration of the industrial goods themselves. To overcome this limitation, we have developed DNP apparatus for the iMATERIA diffractometer, with which we vary the coherent scattering length of hydrogen, by polarizing its nuclear spin via exchange with a paramagnetic center (radical) under extreme conditions of low temperature and high magnetic field. At low temperature 1.2 K and 7 tesla, the electron spin is 95% polarized at thermal equilibrium, whereas that of hydrogen is only 0.3%. This is because of the smaller nuclear magnetic moment. DNP is a technique to transfer the electronic spin state to the hydrogen nucleus using microwave irradiation [38]. For DNP, four technical requirements should be achieved, (1) high magnetic field, (2) low temperature, (3) microwave irradiation-all of which are facility efforts-and (4) radical doping of the sample, which will be a user effort. Figure 17 shows the superconducting magnet with transverse magnetic field, up to 7 tesla and at 1.2 K, prepared for use on iMATERIA (BL20) [39]. The scattering angles of up to 20 • are available for both forward and backward directions along the beam line, which corresponds to 50% of the total area of small-angle and low-angle detector banks. Note that the wavelength-dispersive TOF method overcomes this limitation because it scans q-value by wave length. Homogeneity in magnetic field is of order of 10 −4 in the sample area (ϕ25 mm × 10 mm). The magnet is contained within a vacuum vessel (708 mm diameter) in order to set it in vacuum. It has 4 aluminum windows (0.5-mm thickness) to minimize parasitic scattering. The DNP apparatus is also available for powder diffraction. We change sample specimens using a top-loading stick, on the top of which we set two film specimens and the NMR pick-up coil is also allocated. With this magnet (7 tesla and 1.2 K), we successfully polarized the hydrogen nuclei of polystyrene to P H = 95% [39]. For multicomponent systems, the total small-angle scattering I(q) is given by sum of individual partial scattering functions S ij (q), where the first and second terms are self-and cross-terms, respectively. Relative balance of partial scattering functions is determined by square of the difference in coherent scattering length density ∆b 2 . By performing contrast variation experiments with DNP, we obtain a series of SANS data for a single sample specimen. Then we solve the simultaneous equations to obtain the partial scattering functions. The scattering length depends on polarizations for both neutron and nucleus (P N and P H ).
where P N and P H are polarizations of the neutron and proton respectively. Both of P N and P H affect scattering length b. With P H = 0, which is the case of conventional contrast variation, using isotopic H and D, there is a large difference in b between hydrogen and deuterium. In the case of DNP, on the other hand, P H is controlled so as not to equal to 0. DNP also controls incoherent scattering cross-section (Σ inc ), For positive polarization P H > 0, Σ inc decreases and at P H = P N = 1, it becomes zero. The decrease of Σ inc may be useful for powder diffraction from organic crystalline materials. We mentioned DNP for contrast variation is already open for industrial use at the iMATERIA diffractometer. The detailed analysis of DNP SANS on tire materials is described in [32]: a modern automobile tire is a multicomponent system including many additives (fillers of carbon black and SiO 2 particles, catalyst, all in a matrix of SBR rubber) in order to control roll resistance and stopping quality. The aim is to observe each of the bound rubber on the silica particles, the coupling agent and the tetrasulphide group near the surface selectively.

Perspectives for Completing the Multiscale Observation
In the empty space behind the SANS detector (position B in Figure 2), a USANS detector is installed in order to access q min = 10 −4 Å −1 . Simultaneous scattering measurements using ultra-small-angle, small-angle, low-angle and backward detectors will achieve q max /q mix = 10 6 , which covers the whole hierarchy shown in Figure 1. The pinholes, which shape the incident beam for small-angle scattering (2 θ~0 deg) and do not disturb the backward scattering (2θ~175 deg), have yet to be designed and installed. Powder diffraction (2θ~175 deg) will be measured on the specimens with a plate shape and included in our data reduction program in Figure 6. In the case of the TOF method, incoherent scattering from hydrogen, which is strongly affected by inelastic and multiple scattering, must be quantitatively subtracted. In addition, structural analysis in real space should be quantitatively combined with the SANS analysis [40]. The scattering contrast imaging [41,42] will be useful for industrial research, which is also our future work. Finally, the advanced SANS activities, for example, planned at the European spallation source (ESS) [43,44] will inspire researchers in the small-angle scattering community.

Conclusions
We have established SANS on the powder diffractometer iMATERIA, at BL20 of MLF, J-PARC. The multiscale from Ångstroms to 100 nm, which is described by q max /q mix = 1000, is covered simultaneously by combining the wavelength-dispersive and angle-dispersive methods. To extend applicable materials, we prepared the instrument options of neutron reflectivity and DNP for contrast variation, which are used for industry use.