Implementation and Applications of a Precision Weak-Field Sample Environment for Polarized Neutron Reflectometry at J-PARC

Abstract

Polarized neutron reflectometry (PNR) analyzes surface and interfacial structures of materials. For the SHARAKU reflectometer at the Materials and Life Science Experimental Facility in the Japan Proton Accelerator Research Complex, precise measurements under weak magnetic fields, which are critical for modern spintronics, have long been challenging. To address this issue, we developed a precise weak-field sample environment equipped with a newly designed coil system. The magnetic field at the sample position can be applied within the surface/interface plane, either in the scattering plane (horizontal configuration) or perpendicular to it (vertical configuration). The horizontal configuration achieved high polarization efficiency across a stable field range, whereas the vertical configuration enabled the experiments to cross zero into negative fields. We demonstrated the instrument’s capability by resolving the remanent magnetic structure of an Fe film. Its applicability to soft matter was proven through analysis of a cellulose thin film with roughness using magnetic contrast variation PNR. In this case, precise weak-field control is essential to tune the magnetic contrast from the reference layer beneath the soft film. These results establish the system as a versatile platform for future PNR and polarized off-specular scattering experiments across a wide range of materials.

1. Introduction

Polarized neutron reflectometry (PNR) and polarized off-specular scattering (polarized OSS) are powerful techniques for the high-precision, nondestructive characterization of surface and interface structures in thin films and multilayer materials. Because neutrons are sensitive to magnetic structures through their spin, PNR can nondestructively probe buried magnetic interfaces in systems with high spatial resolution. By contrast, polarized OSS enables the direct observation of domain spacing and correlation lengths on the micrometer scale, providing structural information complementary to that obtained by means of PNR. This capability is particularly important in fields such as high-density magnetic storage and spintronics, where a precise understanding of interfacial magnetism is essential for optimizing device performance. At the Materials and Life Science Experimental Facility (MLF) in the Japan Proton Accelerator Research Complex (J-PARC), the polarized neutron reflectometer SHARAKU (BL17) plays a central role in the study of magnetic thin films. Since its commissioning, BL17 has operated primarily under magnetic fields ranging from 50 Oe to 65 kOe, enabling the detailed characterization of a wide variety of magnetic materials. These measurements have yielded numerous significant findings and have contributed substantially to advances in magnetism and materials science [1,2,3,4].

Although strong-field environments have enabled such progress, an increasing number of studies require precise measurements in the weak-field regime. In recent years, the demand for the analysis of magnetic structures under weak magnetic fields has significantly increased, driven by advances in magnetic and spintronic devices. Precise characterization in the field range of ~10–50 Oe is particularly important for understanding the remanent state and interlayer coupling phenomena relevant to device operation. For example, high-speed magnetic memories operating in the GHz range rely on precise magnetic control near ±30 Oe, and their performance is strongly influenced by the interfacial quality of pseudo-antiferromagnetic multilayers [5,6]. Similarly, magnetic multilayer mirrors used as neutron optical elements have been investigated using PNR and polarized OSS at ~17 Oe to study interlayer exchange coupling [7,8,9,10]. Other materials, including the skyrmionic compound Mn₄N [11], films exhibiting an anomalous Nernst effect [12], and Fe/MgO multilayers [12], exhibit magnetic properties that emerge only in weak-field regimes. Ferromagnetic/ferroelectric heterostructures have also been investigated under combined electric and weak magnetic fields, highlighting the importance of precise magnetic-field control [13]. In addition to these examples, many other studies have reported magnetic behaviors that remain unsaturated or emerge specifically under low magnetic fields. Relevant examples include noncollinear magnetic states [14,15,16]. Furthermore, precise control of weak magnetic fields is essential in modern spintronic devices. Magnetic switching in giant-magnetoresistance and related multilayer systems typically occurs within only a few tens of oersteds, and the coercivity of technologically relevant magnetic films is often below 50 Oe. In this low-field regime, conventional laboratory magnets exhibit large remanent fields and limited positioning accuracy, making it difficult to set the field with the required precision. A key advantage of the WFSE over such strong-field environments is its ability to generate stable and accurately tunable magnetic fields down to 0.4 Oe without field-induced perturbation. This capability enables reliable access to true low-field and remanent states, which is essential for characterizing advanced spintronic materials and magnetic thin-film systems whose functional properties emerge only under weak magnetic fields.

Despite these advances and growing demand, conventional measurement environments often lack the magnetic field resolution, uniformity, and flexibility required to investigate low-field magnetic phenomena. This limitation has impeded the development of next-generation magnetic devices and fundamental understanding of magnetism under weak-field conditions. To overcome the limitations of conventional magnetic field environments, a dedicated sample environment capable of precisely controlling weak magnetic fields must be developed, thereby meeting the growing demands of materials research in this area.

However, only a few widely accessible sample environments can provide stable and precisely controllable magnetic fields in the aforementioned weak-field range. For example, the SuperADAM reflectometer at the Institute Laue–Langevin in France supports sample magnetic fields of ~10 Oe [17]. Similarly, the Platypus instrument at the Australian Nuclear Science and Technology Organization provides an AC magnetic field of ~6 Oe for combined ferromagnetic resonance and PNR experiments. However, these setups are not intended for general-purpose applications such as temperature-dependent or static magnetic field measurements [18].

In this study, we developed a precise weak-field sample environment (WFSE) integrated into the SHARAKU reflectometer. Here, “precision” refers to its experimentally verified capabilities: (i) temporal stability of ~0.2 Oe, (ii) a linear coil current–field relationship, and (iii) spatial uniformity within ~1.1% at the sample position. Unless otherwise noted, we use the abbreviation WFSE. The WFSE supports horizontal and vertical configurations, ensuring a smooth magnetic connection to the guide field, while maintaining a high polarization efficiency for the PNR. The system enables precise and stable control of sample magnetic fields from 0.4 to 125.9 Oe, a range that is challenging to achieve with conventional general-purpose instruments. Furthermore, the system maintained a high neutron polarization efficiency throughout the field range, thereby ensuring its suitability for advanced PNR experiments. This design allows flexible switching between horizontal and vertical configurations, thereby extending the accessible experimental parameter space.

In this study, the terms “horizontal configuration” and “vertical configuration” are mainly used; they correspond to the longitudinal (‖ beam) and transverse (⊥ beam) geometries commonly employed in magnetic scattering. Although this study primarily reports the results obtained using a horizontal field configuration, a vertical configuration is also available for polarized experiments that require continuous measurements across zero-to-negative field conditions. The following sections present the design and performance of the system, demonstrating how its capabilities open up two distinct and powerful research avenues: (i) the elucidation of complex magnetic thin films in the weak-field regime and (ii) the high-precision structural analysis of traditionally nonmagnetic materials, such as soft matter, by utilizing magnetic contrast variation.

To demonstrate the latter avenue, we investigated the applicability of the developed WFSE to systems in which conventional neutron reflectometry encounters difficulties owing to low scattering contrast or significant surface roughness. Magnetic contrast variation PNR (MCV-PNR) [19,20] provides a powerful approach for analyzing thin films when isotopic contrast variation is impractical. In soft-matter and polymer systems, the nuclear SLD contrast between the film and the substrate is often too small to resolve layer structures, and surface or interfacial roughness can suppress Kiessig fringes entirely. By introducing magnetic contrast through a ferromagnetic reference layer, MCV-PNR enables the structural analysis of such low-contrast systems, overcoming limitations that cannot be addressed by standard neutron reflectometry.

In this work, we employed MCV-PNR to determine the nanoscale structure of cellu-lose thin films with high precision. Although cellulose has been widely investigated, its surface and interfacial properties continue to be of significant scientific interest. For the cellulose-based thin-film sample examined here, the surface roughness was too large for conventional neutron reflectometry to resolve the structural features. The use of a magnetic reference layer, together with the precisely controlled weak-field environment provided by the WFSE, enabled reliable extraction of the nanostructural information. The combination of MCV-PNR and a weak-field sample environment thus represents a promising approach for probing the internal structure of soft-matter systems that are difficult to analyze using conventional reflectometry techniques.

These results highlight the potential of the developed WFSE for characterizing magnetic materials under operationally relevant conditions and for precisely determining the nanostructures of soft matter systems. This approach is expected to contribute to advances in a wide range of fields spanning spintronics to sustainable materials science. In the following sections, we present the design and performance of a weak-field sample environment and demonstrate experiments on magnetic thin films and cellulose-based soft materials.

2. Instrumentation

2.1. Design and Performance of the Precision WFSE

This section describes the key components of the BL17 polarized neutron reflectometer [21], newly developed WFSE, and methods used to evaluate its polarization performance. To enable PNR experiments under low magnetic fields, we developed a dedicated WFSE compatible with the BL17 reflectometer at J-PARC in Ibaraki, Japan. BL17 is a time-of-flight (TOF) reflectometer designed for pulsed neutron beams with a wavelength range of 0.24–0.88 nm [22,23]. Its wide wavelength coverage and high polarization performance make it particularly suitable for detailed structural studies of thin films using polarized-neutron techniques. The detailed specifications of the beamline have been reported elsewhere. This study focused on the development and integration of the WFSE into the BL17 system.

The standard 10 kOe magnet installed at BL17 struggles with iron-core remanence, making stable fields within ±50 Oe difficult to attain. A photograph and further details are provided in the Supplementary Material (Figure S1). These strong-field capabilities are invaluable for ferromagnets, whereas weak-field conditions addressed by the WFSE are critical for antiferromagnets and techniques such as MCV-PNR.

To meet the experimental requirements for low-field PNR, the WFSE was designed to (i) generate fields up to ±120 Oe without remanent effects, (ii) ensure a smooth magnetic connection to the guide field, (iii) maintain temporal stability, and (iv) provide high spatial uniformity over a 10 mm × 10 mm sample area. A Helmholtz coil (HC) configuration was adopted to avoid remanent magnetization: two identical circular coils on a common axis, separated by the coil radius, and driven with equal currents. This arrangement yields a uniform field in the central region and a linear response to the applied currents. For magnetic continuity with the beamline, existing guide coils were tuned to operate in a weak-field regime. No magnetic shielding was employed to preserve magnetic continuity with the guide field and maintain experimental access. Heat dissipation was handled by heat sinks and forced-air cooling. The mechanical interfaces between the WFSE and BL17 sample stages were precisely matched to ensure compatibility. The sample was aligned with the geometric center of the coil assembly, enabling accurate positioning within a uniform magnetic field region. We did not adopt a permanently installed measurement device because of its susceptibility to radiation damage.

The spatial uniformity and stability of the magnetic field were evaluated through finite-element simulations using the CAE software Femtet 2019.0 (Murata Software Co., Ltd., Yokohama, Japan). The calculated field variation within a 60 mm × 60 mm × 60 mm volume centered at the sample position was within 1.1%, exceeding the homogeneity required for typical PNR samples. Along the beam axis (z-direction), the field decreased smoothly and crossed zero ~156 mm from the coil center, defining the effective field boundary. Temporal stability was ensured by the high stability of the power supply (±1 × 10⁻⁴ at 24 kW), corresponding to a fluctuation of only ±0.02 Oe, and by maintaining a constant ambient temperature through the facility’s air-conditioning system. These characteristics enable reliable and precisely controlled weak-field conditions necessary for polarized-neutron experiments. The operational range was set to −120 to +120 Oe. To preserve neutron polarization throughout the sample region, the field gradients between the WFSE and surrounding guide coils were minimized by matching their settings to the WFSE conditions.

The WFSE supports both room temperature and cryogenic environments. At room temperature, a non-magnetic sample stick maintains magnetic continuity and suppresses spurious signals, and up to three samples can be mounted using an automated exchanger. For cryogenics, a closed-cycle 4 K cryostat accommodating up to two samples can be installed, enabling sample exchange under cryogenic conditions and enhancing throughput.

A schematic overview of the BL17 beamline is presented in Figure 1, illustrating the layout of the key components used for PNR. The optical system included a supermirror polarizer, two spin flippers, a sample stage, a supermirror analyzer, and a two-dimensional detector [24] arranged along the beam path. The newly developed WFSE was integrated at the sample position, enabling weak-field measurements without compromising neutron polarization. This configuration supported four spin-dependent reflectivity channels through selective spin manipulation before and after the sample.

Figure 2 illustrates two representative WFSE configurations. In the horizontal configuration (Figure 2a), the magnetic field is applied along the beam axis (z-direction). Because the field direction must remain smoothly connected to the upstream and downstream guide coils, polarized-neutron measurements can be performed only under positive applied fields. Applying a negative sample field breaks the magnetic connection to the guide coils, which destroys the neutron polarization and prevents polarized measurements. Negative fields can still be used beforehand to prepare a remanent magnetic state, but PNR data acquisition itself is restricted to positive fields. The usable sample area is approximately 30 mm × 30 mm, defined by the geometry of the sample holder. In the vertical configuration (Figure 2b), the applied field is oriented perpendicular to the beam, and continuous measurements across both negative and positive fields are possible. This allows full characterization of systems requiring field reversal, such as exchange-bias structures. However, at very weak fields (|H| ≤ 10 Oe), field contamination from the guide-field coils and the geomagnetic field can slightly rotate the effective field direction, an effect that must be considered when interpreting low-field PNR data. If such a deviation is acceptable for the experiment, polarized measurements can be conducted at fields as low as 1.4 Oe. A more detailed explanation of the vertical configuration, as well as additional technical information on the WFSE design and performance, is provided in the Supplementary Materials.

To perform the PNR experiments, a well-defined neutron spin state must be prepared and analyzed. Polarized neutrons are generated by passing unpolarized neutrons through a supermirror polarizer that selects spin states aligned antiparallel to the guide magnetic field. After interaction with the sample, the spin state of the reflected neutrons was analyzed using a supermirror analyzer. Spin-state manipulation was achieved by placing two spin flippers upstream and downstream of the sample. By activating these flippers in various combinations, all four spin-dependent reflectivity channels (R⁺⁺, R⁻⁻, R⁺⁻, and R⁻⁺) can be accessed. The performances of the polarizer and analyzer were characterized by their polarization efficiencies P₁ and A₂, respectively. These are defined as follows:

$$P={\displaystyle \frac{I^{+}-I^{-}}{I^{+}+I^{-}}}$$

(1)

where I⁺ and I⁻ are the intensities of neutrons with spin up and spin down, respectively, at the position of the polarizer or analyzer. These efficiency values are essential for correcting the measured reflectivity and extracting accurate spin-dependent signals.

The spin-manipulation system at BL17 includes two types of spin flippers: a two-coil flipper [25] located upstream of the sample and a Mezei-type flipper [26] positioned downstream. These devices are part of the standard optical configuration of the beamline and have been described in detail elsewhere [21]. Owing to the finite flipping efficiencies, the actual neutron spin states differ slightly from the ideal configurations. To characterize the overall spin-selective performance of the system, the combined polarization efficiency, P_total, is defined as the product of the polarizer, spin-flipper, and analyzer efficiencies:

$$P_{total}=F_1·F_2·\left|P_1\right|·A_2$$

(2)

where P₁ and A₂ are the polarization efficiencies of the polarizer and analyzer, respectively; and F₁, F₂ represent the flipping efficiencies of the upstream and downstream spin flippers, respectively.

In the practical measurements, spin-dependent reflectivity signals were obtained by operating the upstream and downstream spin flippers in four on/off combinations. The resulting neutron intensities are denoted as I(ij), where i and j indicate the flipper states (0 = off and 1 = on) upstream and downstream of the sample, respectively. Each I(ij) configuration primarily corresponds to a dominant spin-state combination.

Table 1. Operational states of the spin flippers and the corresponding measured spin-dependent reflectivity channels. “off” indicates no spin flip, and “on” indicates spin inversion. Owing to the finite polarization and flipping efficiencies, the spin states shown represent the dominant components in each configuration.

Upstream Flipper	Downstream Flipper	I(ij)	Corrected Intensity
off (0)	off (0)	I0(00), I(00)	I0−+, R−+
off (0)	on (1)	I0(01), I(01)	I0−−, R−−
on (1)	off (0)	I0(10), I(10)	I0++, R++
on (1)	on (1)	I0(11), I(11)	I0−+, R−+

summarizes the relationship between the flipper settings and corresponding measured spin-dependent reflectivity channels.

Owing to the finite polarization and flipping efficiencies of the optical components, the measured intensities I(ij) are not pure representations of a single spin state, but contain contributions from other channels. To extract accurate spin-dependent reflectivities, the measured signals must be corrected by accounting for the polarization efficiencies of the polarizer and analyzer (P₁, A₂) as well as the flipping efficiencies of the upstream and downstream flippers (F₁, F₂). This correction is essential for quantitative analyses, particularly when calculating spin asymmetries or fitting structural models to reflectivity data. The relationship between the measured intensities and the true spin-dependent reflectivities can be expressed using a correction matrix that accounts for the finite polarization and flipping efficiencies of the optical components. This matrix describes how each measured intensity I(ij) is a linear combination of the ideal reflectance R⁺⁺, R⁻⁻, R⁺⁻, and R⁻⁺ weighted by the performance of the polarizer, analyzer, and spin flippers. In the subsequent data analysis, corrections for the polarization and flipping efficiencies of the polarizer, analyzer, and spin flippers were applied to obtain the true spin-dependent reflectivities. These corrections are essential for accurately extracting magnetic signals, particularly under weak magnetic fields.

To evaluate the practical performance of the newly developed WFSE, we conducted a series of measurements focusing on the magnetic field calibration and polarization efficiency. The magnetic field was measured at the sample position using a Gaussmeter (Model 5180, F.W. Bell, Milwaukie, OR, USA), with the probe rigidly mounted on a fixed stand and positioned at the center of the HC. Figure 3 shows the calibration of the on-sample field versus the HC current. The temporal fluctuation during acquisition was ~0.2 Oe (20 µT); the corresponding error bars (±1 s.d.) are smaller than the symbol size.

To quantify the polarization performance, we performed four-channel direct-beam and reflection measurements using a magnetite (Fe₃O₄) single crystal magnetized to saturation. The overall polarization efficiency of the beamline, including the spin flippers, polarizer, and analyzer, was assessed in the horizontal configuration of the WFSE. For reference, these results were compared with the polarization efficiencies obtained using a conventional 10 kOe magnet. The individual polarization efficiencies of the upstream and downstream spin flippers (F₁, F₂), polarizer (P₁), and analyzer (A₂) were extracted by solving a set of equations based on the measured intensities, following the matrix-correction formalism summarized in the Supplementary Materials (Equations (S3)–(S7)). To examine the magnetic field dependence of the polarization efficiency, measurements were carried out across a range of sample magnetic fields from +0.4 to +125.9 Oe using the horizontal configuration of the WFSE. Polarization efficiency measurements were performed using a 10 mm × 10 mm magnetite (Fe₃O₄) single crystal with a (100) surface orientation, grown using the floating zone method (SurfaceNet GmbH, Rheine, Germany). The proton beam power at the mercury target at the MLF at J-PARC was 650 kW. The neutron incidence angle was set to 0.7°, and the angular divergence was defined as 5.0% using the slit settings. For each magnetic field condition, the total measurement time for the four-channel direct was 1.5 h. The four-channel reflection spectra for the polarization efficiency determination were acquired at an incident angle of 0.7° with a measurement time of 4 h, avoiding the total reflection region.

2.2. Magnetic-Field Characteristics and Polarization Efficiencies

This section presents the key results and their implications using the WFSE: (i) calibration and stability of the on-sample field, (ii) direct-beam polarization spectra, (iii) wavelength dependence of component efficiencies, and (iv) field dependence of the total efficiency.

Figure 3 shows the on-sample field versus HC current in the horizontal WFSE configuration. The response is linear:

H (Oe) = 6.226 × I (A) + 1.4

(3)

where the +1.4 Oe offset arises from the guide and geomagnetic fields (each +0.7 Oe). For example, I = 20 A yields H = 125.9 Oe, whereas I = 0 A yields H = 1.4 Oe. The temporal fluctuation during acquisition was ~0.2 Oe (±1 s.d.), which defines the practical floor of field control and corresponds to ~±0.03 A in the HC current. The specified current noise of the HC power supply (±30 A, ±70 V, 0.01% current noise; Model TM-PSBC3070-234-S, Kudo-Denki, Kashiwa, Japan) is sufficiently low that the observed ~0.2 Oe fluctuation reflects combined contributions from electronics, D/A conversion, and ambient magnetic noise inside and outside the beamline. These results demonstrate that the WFSE delivers stable, precisely tunable weak fields over the full operating range.

Figure 4 shows the four-channel direct spectra acquired in the horizontal configuration at +1.4 Oe (0 A in the HC). The usable wavelength band (0.24–0.88 nm) was determined by the source-chopper setting. All spin-resolved spectra varied smoothly with the wavelength, confirming a stable source and polarization optics. The residual signals in I₀(00) and I₀(11) were small yet finite, attributable to optical imperfections and the background.

Figure 5 summarizes the wavelength dependence of the polarization efficiencies extracted from the four-channel data. The polarization efficiencies of F₁, F₂, and P₁ remained nearly constant throughout the wavelength range of 0.24–0.88 nm. Conversely, A₂ and P_total gradually decreased toward longer wavelengths. Quantitatively, A₂ dropped from ~99% (0.25 nm) to 87% (0.87 nm); P_total declines from ~95% to ~88% toward longer wavelengths. We attribute this to a geometric mismatch between the sheet-beam and the fixed analyzer orientation optimized for 2θ-axis detection in polarized OSS mode. Practical mitigation without hardware changes involves weight acquisition toward shorter wavelengths via chopper phase optimization. Despite this reduction, the overall polarization performance remains sufficiently high across the full wavelength range, ensuring the reliability of the weak-field PNR measurements.

Figure 6 illustrates how P_total varies with the applied magnetic field for selected wavelengths in the horizontal WFSE configuration. The total polarization efficiency, P_total, is essentially independent of the field at all tested wavelengths from 0.4 to 125.9 Oe. The lower measurement bound of 0.4 Oe reflects the ~0.2 Oe calibration fluctuation; below this level, we did not attempt a quantitative efficiency evaluation. This field-independent behavior is consistent with the magnetic-connection design of the WFSE.

In the horizontal configuration, the WFSE field is parallel to the guide-field direction, and its magnitude is not large enough to disturb the upstream and downstream polarization devices. As long as the polarization efficiency remains unchanged between the minimum and maximum applied fields, it can be assumed to remain constant at intermediate fields as well—an observation directly confirmed in Figure 6.

For the vertical configuration, neutron-spin transport was evaluated using the spatial field distribution obtained from the Femtet simulations. As detailed in Section S2 of the Supplementary Materials, the WFSE was designed such that the ratio of the neutron Larmor precession frequency ω_L to the magnetic-field-variation rate ω_B exceeds 10 under adiabatic conditions, ensuring that the neutron spins follow the field direction and that the polarization efficiency is preserved across the entire operating field range.

The performance characteristics of the WFSE presented in this section establish a reliable foundation for the experiments on the magnetic thin-film in the remanent state and the structural analysis of cellulose-based soft matter samples using magnetic contrast variation PNR (MCV-PNR), which are presented in the following sections.

3. Applications

3.1. PNR on Fe Thin Film in the Remanent State

3.1.1. Method

To investigate the magnetic structure in the remanent state of the sample, PNR measurements were performed using SiO₂ (100 nm)/Fe (20 nm)/MgO. The field-dependent magnetization of the Fe layer was measured using SQUID magnetometry (MPMS II, Quantum Design, Tokyo, Japan). An external field was applied parallel to the film plane. The resulting M–H curve was used to identify the remanent state in the PNR experiments.

To determine the depth-resolved magnetic structure in the remanent state, we performed four-channel PNR using WFSE in the horizontal configuration. The sample was first magnetized at −120 Oe and then measured at +10 Oe, where the M–H curve indicated proximity to remanence. Magnetic structure analysis was performed using four-channel and two-channel PNR. Data acquisition and processing were performed using a system developed at the MLF [27]. The neutron angular divergence $d\theta /\theta$ was set to 5.0%, as defined by the slit settings. Data was reduced using Utsusemi 4.0.0 [28] with a TOF binning width of 1.0 ms, corresponding to a wavelength bandwidth of 0.02 nm. To optimize the counting statistics, the measurement times were allocated in proportion to the signal intensity: I(00):I(01):I(10):I(11) = 6:1:1:6, because I(01) and I(10) exhibited higher intensities than I(00) and I(11). The reflectivity data were structurally fitted using the GenX 3.0 software package [29]. To map the field dependence of the average moment with higher throughput, we acquired two-channel PNR (non-analyzed, R⁺ and R⁻) from 10.0 to 66.0 Oe. The spin asymmetry SA = (R⁺ − R⁻)/(R⁺ + R⁻) was used as an indicator sensitive to magnetization.

3.1.2. Analysis and Results for Fe Thin Film in the Remanent State

The WFSE provides a significant advantage for polarized-neutron measurements in the remanent state. As noted in the introduction, conventional strong-field magnets exhibit substantial remanent fields, making true zero-field or controlled low-field conditions difficult to achieve. In contrast, the WFSE delivers stable and accurately tunable fields down to 0.4 Oe, enabling the intrinsic magnetic-domain configuration in the remanent state to be probed without field-induced perturbation. This capability is essential for studying domain formation, magnetization canting, and depth-dependent magnetic structures in soft magnetic thin films.

Figure 7a shows the magnetization curve of the SiO₂ (100 nm)/Fe (20 nm)/MgO sample. To investigate the magnetic structure under the remanent state, four-channel PNR spectra were obtained at +10 Oe after presetting the magnetic history at −120 Oe, as shown in Figure 7b. Figure 7c shows the nuclear and magnetic scattering length–density profiles, nSLD and mSLD, extracted from the fits. The mSLD profile indicates that the Fe layer exhibits a magnetic moment of 1.34 μ_B ± 0.06 per atom, with a tilt angle of 86° ± 4° relative to the external field direction. The negative sign of the mSLD reflects that the average magnetization vector is oriented opposite to the neutron spin-quantization axis. It should be emphasized that the magnetic moment obtained from PNR represents a spatial average over the entire illuminated area—typically on the order of square centimeters—rather than the magnetization of a single domain. In the remanent state, the Fe film likely breaks into multiple in-plane domains with magnetization directions distributed among several easy axes. Vector cancellation among these domains naturally yields a net moment smaller than the saturation value of ~2.2 μ_B. To interpret the magnetic configuration more quantitatively, we evaluated the moment tilt angle within an effective single-domain model, where the angle is defined symmetrically with respect to the external field in the range from 0° to 180°. As evident in Figure 7b, the spin-flip reflectivities in the total-reflection region become comparable to R⁺⁺ and R⁻⁻, indicating that the magnetic moments are oriented approximately 90° relative to the applied field. If the domain moments are randomly distributed, the average over the 0–180° range naturally yields a tilt angle of 90°, whereas full saturation along the field direction would result in a tilt angle close to 0°. This explains why the experimentally obtained tilt angle appears near 90° in the remanent state. Figure 7d–g present the two-channel PNR spectra measured under various weak-field conditions, demonstrating the evolution of the magnetic moment following different field histories.

The WFSE developed in this study significantly expanded the measurement capabilities of the BL17 reflectometer. In particular, it enables PNR in a previously inaccessible field range of 0.4–125.9 Oe, which could not be achieved using the conventional 10 kOe magnet owing to its intrinsic remanence. This advancement allows for a detailed investigation of magnetic phenomena, such as the remanent state and domain formation, under near-zero field conditions. In addition, the results demonstrate that WFSE enables a detailed, nondestructive analysis of magnetic thin films in remanent states, a regime that is highly relevant to practical device operation. In particular, such measurements are essential for understanding the magnetic domain behavior and interlayer coupling in complex multilayer structures, such as the pseudo-antiferromagnetic stacks used in spintronic devices. By enabling the precise evaluation of the magnetization direction and moment magnitude under low-field conditions, this environment provides critical insights into magnetic switching mechanisms, domain stability, and performance tuning in next-generation magnetic memory applications. The broader significance of these findings and future research directions are summarized in the conclusion.

To characterize magnetic thin films whose functional properties depend sensitively on their zero-field magnetic configuration, it is essential to perform measurements under well-defined remanent-state conditions. The WFSE enables such measurements by providing a stable and accurately controlled weak-field environment, allowing polarized neutron reflectometry (PNR) and off-specular scattering (OSS) to be conducted without applying an external magnetic field. PNR yields depth-resolved information on the magnetic moment and interfacial roughness, while OSS provides statistical in-plane roughness parameters such as the lateral correlation length and Hurst exponent. This capability is particularly advantageous for multilayer magnetic systems, where magnetoresistive properties often peak at weak applied fields on the order of several tens of oersteds. The WFSE thus allows systematic investigation of zero-field, antiferromagnetic, and ferromagnetic remanent states in technologically relevant magnetic heterostructures.

3.2. Structural Analysis of Cellulose Acetate (CA) Film with Substantial Surface Roughness Using Magnetic Contrast Variation PNR (MCV-PNR)

3.2.1. MCV-PNR Method

This section introduces the MCV-PNR technique [19,20], with a focus on its application in the structural analysis of cellulose-based soft matter samples. As an application example, we conducted MCV-PNR measurements on a cellulose thin-film/Fe/Si sample to demonstrate the capability of determining the nanostructure of cellulose films with high precision [30]. MCV-PNR is a valuable technique for accurately analyzing the surface, interface, and thin-layer structures of cellulose thin films. As a preliminary step, we measured the neutron reflectivity (NR) of a cellulose acetate (CA) thin film, ~150 nm thick, prepared on a quartz substrate (ESI-01), as shown in Figure 8a. However, the NR data did not exhibit any interference fringes from the CA thin film, probably owing to its significant surface roughness. Because MCV-PNR is particularly well suited for analyzing thin films with substantial surface roughness, we applied this method to investigate the CA thin-film structure.

Iron has a large magnetic scattering length, making it an excellent material for MCV-PNRs. Therefore, in this study, we prepared CA/Fe/Si samples in which an iron layer was incorporated as a magnetic contrast medium. The Fe/Si substrate was fabricated using a sputtering system at the AIST Nano-Processing Facility. The CA layer was prepared by spin coating a 2 wt% cellulose acetate/acetone solution at 2000 rpm using a spin coater (MS-A150, Mikasa, Tokyo, Japan). The sample was then heated at 90 °C in a vacuum oven for 12 h and subsequently stored in a box under low-humidity and dust-free conditions at room temperature. Prior to the neutron experiments, the magnetization curve of the Fe layer was measured using SQUID magnetometry and was found to saturate at ~30.0 Oe. Consequently, all MCV-PNR measurements were performed under a static field of 60.0 Oe to guarantee full saturation of the magnetic reference layer using the horizontal configuration of the WFSE. For samples with thicker films, observing Kiessig fringes in the low-Q_z region was preferable during the reflectivity measurements. In the CA/Fe/Si sample, the back-incidence geometry, in which neutrons are introduced through the substrate, allows access to lower momentum-transfer regions (Q_z), which is advantageous for resolving Kiessig fringes in thicker films. Therefore, MCV-PNR measurements were conducted in back-incidence mode to obtain high-quality data suitable for analyzing the CA/Fe/Si sample. A beam footprint of 10 mm × 10 mm was maintained on the sample surface, and TOF neutron data were collected using a two-dimensional scintillation detector with spatial resolution capabilities. All measurements were performed at ambient temperature under a magnetic field of 60.0 Oe, and magnetization of the iron thin film was confirmed to be fully saturated. The MCV-PNR profiles were fitted using Motofit software with Igor pro 8 [31,32] with a least-squares approach to minimize fitting deviations. The film thickness (t, Å), scattering length density (SLD, ρ), and Gaussian interfacial roughness (σ, Å) were evaluated using Motofit.

3.2.2. Analysis and Results Using MCV-PNR

To evaluate the applicability of WFSE to soft-matter systems, we employed the MCV-PNR method to analyze the CA/Fe/Si samples. Owing to its substantial surface roughness, this sample was not suitable for structural analysis using conventional neutron reflectometry, as indicated by the lack of interference fringes in the NR data (Figure 8a).

To overcome this limitation, we introduced a magnetic reference layer beneath the CA film to generate a spin-dependent contrast. The magnetization curve of the sample is shown in Figure 8b. A magnetic field of 60 Oe was applied to saturate the magnetization of the Fe reference layer. Under these conditions, the contrast in the R⁻ channel (nSLD–mSLD) is enhanced, clearly revealing the Kiessig fringes. Figure 8c shows the MCV-PNR reflectivity profiles of the CA/Fe/Si sample measured under a magnetic field of 60 Oe. A significant contrast between the R⁺ and R⁻ reflectivity curves was observed, resulting from the magnetic scattering contribution of the Fe layer. Here, R⁺ and R⁻ represent the reflectivities for neutrons incident on the sample with their spins parallel and antiparallel to the applied magnetic field, respectively. In the reflectivity model, R⁺ corresponds to (nSLD + mSLD) and R⁻ to (nSLD − mSLD), where nSLD is the nuclear scattering length density and mSLD is the magnetic one. Kiessig fringes were apparent in the R⁻ profile but not in the R⁺ profile, indicating that the contrast in scattering length density (nSLD − mSLD) between the CA and Fe layers is more favorable under the R⁻ condition. These results confirm that the MCV-PNR method can enhance the structural visibility of soft films with high surface roughness, which is otherwise difficult to resolve using conventional NR techniques. The reflectivity data were fitted using a model consistent with the known layer structure, yielding reliable nSLD and mSLD profiles, as shown in Figure 8d. The mSLD profile revealed a saturated magnetic moment of 2.2 μ_B per atom in the Fe layer, while the CA layer exhibited negligible magnetization, as expected. No change was observed in the nSLD between the R⁺ and R⁻ datasets, confirming that the observed contrast originated from magnetic rather than nuclear scattering. The surface roughness at the air/CA and CA/Fe interfaces was estimated to be 7.8 nm and 2.6 nm, respectively. These values confirm that the CA layer possessed substantial surface roughness, which hindered structural analysis using conventional neutron reflectometry techniques.

Su et al. reported a surface roughness of 2.8 nm for CA thin films prepared by spin coating with a 2.5 wt% polymer solution [30]. The higher roughness observed in our study was presumed to result from differences in the deposition conditions. Despite this discrepancy, the structural parameters obtained from the MCV-PNR were in close agreement with the design values, underscoring the robustness of the method.

The successful detection of Kiessig fringes in the R⁻ channel despite such conditions demonstrates the practical advantage of MCV-PNR for soft systems that are otherwise challenging to characterize. These findings establish MCV-PNR as a practical tool for the nondestructive structural analysis of soft matter systems under weak magnetic field conditions. Because the magnetization of the Fe reference layer typically saturates at around 30 Oe, precise control of the applied field below this level is essential for tuning the SLD contrast. The ability of the present system to regulate weak magnetic fields with high stability demonstrates its usefulness for a wide range of soft-matter investigations.

4. Conclusions

In conclusion, we developed a WFSE for a BL17 polarized neutron reflectometer at MLF J-PARC. The system enables the precise control of magnetic fields from sub-Oersted to over 100 Oe and maintains a high polarization efficiency. Demonstration experiments confirmed its utility in analyzing magnetic thin films in the remanent state and resolving the structural features of soft matter films using MCV-PNR. This system provides a practical platform for low-field NR of a wide range of materials. Future plans will involve improving the experimental throughput by introducing a GM-type 4 K cryostat and a multisample exchange mechanism. We also plan to adapt the system to support polarized off-specular scattering (polarized OSS) measurements, enabling the observation of more detailed magnetic domain structures, thereby expanding the range of applications of the system.

This system is expected to accelerate the research and development of spin electronics, magnetic recording, and energy-conversion materials. Its ability to control weak magnetic fields with high precision opens new avenues for studying magnetic materials in realistic remanent states. In soft-matter research, this instrument enables nondestructive profiling of rough organic films, offering a powerful route for the structural analysis of complex organic and biological layers.