FULL PAPERS Journal of the Physical Society of Japan Analysis of Multifrequency Oscillating Magnetic Fields by Neutron Spin Interferometry Ryuto Fujitani1,2 * , Masahiro Hino2 , and Takashi Higuchi2,3 1 arXiv:2605.28673v2 [physics.ins-det] 1 Jul 2026 2 Department of Nuclear Engineering, Kyoto University, Kyoto 615-8246, Japan FULL PAPERS J. Phys. Soc. Jpn. Thus, the phase shift Ω(θ) can be rewritten as the neutron, v is the neutron velocity, and l is the distance between RSFs. B(x) is the strength of the magnetic field along the neutron path x between the RSFs. In practice, owing to the finite polarization efficiencies of the polarizer and analyzer, as well as the spin-flip efficiency of RSFs, the contrast C of an interference pattern does not reach unity. Therefore, the experimentally observed interference pattern is described by Ω(θ) = N0 (6) {1 − C cos(χ − P)} , 2 where the phase of an interference pattern is P. Parameters C, P and N0 can be extracted by fitting the measured neutron count to Eq.(6). Theoritical Description This equation describes an integration of the dispersion of the phase shift. Introducing the complex value c defined as Z 2π 1 c= exp (iΩ(θ))dθ, (15) 2π 0 3.1 Formulation of multifrequency oscillating magnetic field A multifrequency oscillating magnetic field B(x, t) at position x and time t can be expressed as a Fourier series expansion based on the fundamental frequency f . B(x, t) = B0 (x) ∞ X (13) 3.2 Formulation of interference pattern Consider an interference pattern obtained for an oscillating magnetic field with continuous exposure. In the case of continuous exposure, the phase θ can take any value within the range 0 ≤ θ ≤ 2π. Therefore, the observed interference pattern corresponds to an average over all the values of θ. Consequently, an observed interference pattern is given by Z 2π N0 1 (14) N= {1 − cos(χ − Ω(θ) − δ)} dθ. 2 2π 0 N= 3. ∞ 2|µn | X |b(nk)|dn sin (nθ + ϕn + θ(nk)) . ℏv n=1 an interference pattern can be writen as dn sin(−2πn f t + ϕn ) (7) N= n=1 Here, ϕn denotes the phase of the n-th frequency component oscillation and dn is the amplitude of the n-th component normalized by the fundamental amplitude (d1 = 1). B0 (x) represents the strength of the magnetic field without time dependence. When a neutron enters a region of an oscillating magnetic field at time τ, in the same way as in Eq. (5), the phase shift Ω observed in spin interference is given by Z ∞ X 2|µn | l dn sin (−2π f n(t − τ) + ϕn ) dx. B0 (x) Ω(τ) = ℏv 0 n=1 (8) Since the magnetic field oscillates while a neutron traverses the field, the magnetic field experienced by a neutron is obtained by substituting t = x/v into Eq. (8). Thus, Eq. (8) can be rewritten as ! Z ∞ X 2|µn | ∞ 2π f Ω(θ) = dn sin − nx + nθ + ϕn dx. B0 (x) ℏv −∞ v n=1 (9) Here, θ represents the phase of the oscillating magnetic field at the moment a neutron enters the field, defined by θ = 2π f τ. By assuming B0 (x) = 0 at x ≤ 0 or x ≥ l, the integration range has been extended to infinity. Equation (9) provides the expression of Ω for an oscillating magnetic field, corresponding to the static case in Eq. (5). Introducing k = 2π f /v The value of c is X c= Jm1 m1 +2m2 +3m3 +···=0 mk ∈Z (16) ! ! 2|µn | 2|µn | |b(k)| Jm2 |b(2k)|d2 · · · ℏv ℏv × exp {i(m1 (ϕ1 + θ(k)) + m2 (ϕ2 + θ(2k)) + · · · )} , (17) where Z denotes the set of all integers and Jm (x) denotes the m-th order Bessel function of the first kind. The derivation of Eq. (17) is as follows. We use the generating function of the Bessel function, given by !! ∞ X u 1 w− = Jm (u)wm , (18) exp 2 w m=−∞ where w and u are arbitrary complex numbers, with w , 0. By multiplying both sides, we obtain !! Y ∞ ∞ X ∞ Y un 1 n exp wn − = Jmn (un )wm (19) n . 2 w n n=1 n=1 m =−∞ n i(nθ+ϕn +θ(nk)) n| and un = 2|µ ℏv |b(nk)|dn , the left-hand Setting wn = e side of Eq. (19) becomes  ∞    X 2|µn |  exp i |b(nk)|dn sin(nθ + ϕn + θ(nk)) = exp (iΩ(θ)) , ℏv n=1 (20) while the right-hand side of Eq. (19) becomes ! ∞ X ∞ Y 2|µn | Jmn |b(nk)|dn exp (imn (nθ + ϕn + θ(nk))) . ℏv n=1 mn =−∞ (21) Integrating both sides with respect to θ from 0 to 2π, and making use of the orthogonality in θ, we obtain Eq. (17). (10) and using the k-space Fourier transform of the magnetic field profile B0 (x), the power spectrum |b(k)| and the phase θ(k) are defined as Z ∞ |b(k)|eiθ(k) = B0 (x)e−ikx dx. (11) −∞ Extracting only the imaginary part of Eq. (11) gives Z ∞ |b(k)| sin(θ(k)) = B0 (x) sin(−kx)dx. N0  1 − |c| cos(χ − δ − arg (c)) . 2 3.3 Case of single-frequency oscillation Consider the case in which a single-frequency oscillation is applied. Under this condition, dn = 0 for n ≥ 2. Using the properties of the Bessel function, J0 (0) = 1 and Jn (0) = 0 for (12) −∞ 2 FULL PAPERS J. Phys. Soc. Jpn. 1HXWURQVV  RSFs, which flip the neutron spin orientation, generate 20 kHz oscillating magnetic fields along the x-axis. The oscillating magnetic field to be analyzed by the spin interferometry is produced by a rectangle coil named the sample coil. Its dimensions are 100 mm in the z-direction (height), 45 mm in the x-direction (length), and 70 mm in the y-direction (width), and it consists of 15 turns. Its wire is 1 mm thick and is made of Al, enabling lower neutron absorption. The spatial distribution of the magnetic field at the center of the coil along the x-axis calculated by using the Biot–Savert law for a current of 1 A is shown in Fig. 3. An interference pattern obtained with this setup is shown in Fig. 1. Fitting the data to Eq. (6), the contrast C of the interference pattern is determined to be 0.9659 ± 0.0077 and the phase P is 33.65 ± 0.64 deg. To demonstrate Eq. (22), we applied a single-frequency oscillating current to the coil. Interference patterns were measured while varying the amplitude of the AC current from 0.0 to 3.0 A peak-to-peak (App ). Measurements were carried out at frequencies from 2500 to 20000 Hz. Subsequently, to demonstrate Eq. (24), we applied an AC current composed of fundamental and second-harmonic oscillations. We measured interference patterns while varying the phase ϕ2 of a double-frequency oscillation shown in Eq. (24) from 0 to 360 deg. The amplitude of AC current for a fundamental oscillation was fixed to 2 App , while the relative amplitude of a second-harmonic oscillation, denoted as d2 , was varied among 0.25, 0.5, 0.75 and 1. Measurements were performed for fundamental frequencies from 2500, 5000, 7500 and 10000 Hz.    Fig. 1. Fig. 2.     3KDVHEHWZHHQ56)V  GHJ Interference pattern obtained from the spin interferometer shown in n , 0, Eq.(17) becomes c = J0 ! 2|µn | |b(k)| . ℏv (22) This result corresponds to the expression previously reported.20) Since c is always real and arg (c) shown in Eq. (16) is always zero, the phase of the interference pattern remains constant, while only the contrast decays following the zero-th order Bessel function. 3.4 Case of fundamental and second-harmonic oscillations Consider fundamental and second-harmonic oscillations given by 5. 5.1 Case of single-frequency oscillation In Fig. 4, the contrast follows the zero-th order Bessel function of the applied current as predicted by Eq. (22). The vertical axis in Fig. 4 represents the contrast of the interference pattern, while the horizontal axis indicates the amplitude of the AC current applied to the coil. Data for each frequency are plotted on the same graph. The plots represent experimental data, which have been scaled by a factor of 1/0.9659, the inverse of the contrast observed without the sample, as shown in Fig. 1. By introducing the power spectrum |b1App (k)| for a current of 1 App and the AC current amplitude I, Eq. (22) can be rewritten as ! 2|µn | |b1App (k)| × I . (25) J0 ℏv B(x, t) = B0 (x) {sin(−2π f t + ϕ1 ) + d2 sin(−4π f t + ϕ2 )} . (23) Introducing dn = 0 for n ≥ 3, Eq. (17) becomes ! ! ∞ X 2|µn | 2|µn | c= J−2m |b(k)| Jm |b(2k)|d2 ℏv ℏv (24) m=−∞ × exp {i(−2m(ϕ1 + θ(k)) + m(ϕ2 + θ(2k)))} . Since c can be a complex value and arg (c) shown in Eq. (16) is not always zero, the phase of the interference pattern can be shifted because of the oscillating magnetic field. 4. Results and Discussion Experimental Setup To demonstrate Eqs. (22) and (24), we conducted an experiment on the JRR-3 C3-1-2-2 (Multilayer Interferometer and reflectometer for NEutron 2, MINE2) beamline. The incident neutron beam is continuous and monochromated to a wavelength of 8.8 Å and the full width at half maximum of 2.7%. The experimental setup is shown in Fig. 2. The beam direction is along the x-axis. The setup of the spin interferometer consists, in order from the upstream, of a polarizer, RSF1, a sample coil, RSF2, an Sanalyzer, and a neutron detector. A guide magnetic field is applied throughout the entire spin interferometer system to maintain the neutron spin quantization axis. The polarizer and analyzer, which reflect only spin-up neutrons, are composed of magnetic multilayer mirrors. The The lines represent fits to Eq. (25). By fitting data to Eq. (25), the value of 2|µn ||b1App (k)|/ℏv can be extracted. Figure 5 shows the values of 2|µn ||b1App (k)|/ℏv as a function of the frequency of the oscillating magnetic field. The line represents the value calculated using the magnetic field shown in Fig. 3. The plots are obtained from the measured results shown in Fig. 4. Since k is proportional to the frequency, the values of 2|µn ||b1App (k)|/ℏv have a frequency dependence. The experimental data reproduce the model well. The remaining discrepancy can be considered to be imperfections in the coil shape. While the calculated values were obtained using the Biot–Savart law, distortions in the coil or its installation angle can affect the value of |b1App (k)|. 3 FULL PAPERS J. Phys. Soc. Jpn. JXLGHFRLO 56) N+] SRODUL]HU 56) N+] DQDO\]HU VDPSOH FRLO ]   Fig. 2.  3RVLWLRQ PP   [  Experimental setup for a neutron spin interferometer at JRR-3 C3-1-2-2.   &DOFXODWLRQ ([SHULPHQWDO   v |b1App (k)|    2| n | 0DJQHWLFILHOG P7 GHWHFWRU \        3RVLWLRQ PP        )UHTXHQF\ +] Fig. 5. Calculated and measured values of 2|µn ||b1App (k)|/ℏv. The calculated values are obtained from B(k) shown in Fig. 3 using the Biot–Savart law and a neutron velocity corresponding to 8.8 Å. These are compared with the measured values shown in Fig. 4. Fig. 3. Magnetic field generated by sample coil when 1 A current is applied. The calculation was performed using the Biot–Savert law.  &RQWUDVW  phase is shown in Fig. 7. These results correspond to the case where the fundamental oscillation of the applied current is 2 App . Here, the frequency in the legends refers to that of a fundamental oscillation. The contrast is scaled by a factor of 1/0.9659. The lines represent the results calculated using Eq. (24) and the measured 2|µn ||b1App (k)|/ℏv in Fig. 5. In the calculation, the summation was performed over the range m = −20 – 20. In the experiments, the AC current was generated by the function generator and a power amplifier. Because of the inductance, determining the phase of a second-harmonic oscillation, ϕ2 , is difficult. Therefore, in Figs. 6 and 7, the experimental plot is shifted horizontally to the calculation results with phase ϕ2 . Both the contrast and phase exhibit periodic behavior with respect to ϕ2 , but they are not analytically simple such as sinusoidal. Despite the complexity of these variations, the experimental results show reasonable reproducibility. For data with the fundamental frequency of 2500 Hz, the contrast and phase shown in Figs. 6 and 7 exhibit the most significant changes among the four frequencies. Changes in the phase of a second-harmonic oscillation depend on the       +] +] +] +] +] +]   $PSOLWXGHRI$&FXUUHQW $SS  Fig. 4. Contrast of interference patterns varying with the amplitude of the sinusoidal AC current applied to the coil. 5.2 Case of fundumental and second-harmonic oscillations The contrast measured as a function of the phase of a second-harmonic oscillation, ϕ2 , is shown in Fig. 6, and the 4 FULL PAPERS J. Phys. Soc. Jpn. D d2   E d2  values of 2|µn ||b1App (k)|/ℏv are smaller at 10000, 15000, and 20000 Hz. Therefore, the data for fundamental frequencies of 5000, 7500, and 10000 Hz show slight changes in contrast and phase. To evaluate the effective frequency range, we consider a rectangular magnetic field B0 (x) whose length is L, frequency is f , and strength is B:    B 0 ≤ x ≤ L (26) B0 (x) =   0 otherwise.   &RQWUDVWRILQWHUIHUHQFHSDWWHUQV    F d2   G d2   From Eq. (11), the following equation can be obtained: ! πfL 2|µn | B 2|µn | sin . (27) |b(k)| = ℏv ℏ πfL v    Equation (27) shows that 2|µn ||b(2k)|/ℏv vanishes at f L/v = 1 and generally decreases as f increases. Thus, the fundamental frequency f should satisfy f < v/L in order to obtain measurable changes. The n-th harmonic also remains observable as long as n f < v/L. Although the magnetic field in this experiment is not perfectly rectangular, as shown in Fig. 3, a rough estimation is sufficient for this study. Given the conditions of L = 45 mm and v = 450 m/s, 2|µn ||b(k)|/ℏv becomes zero at f = 10000 Hz and smaller for f > 10000 Hz. Therefore, the data in Figs. 6 and 7, which have fundamental frequencies of 5000, 7500, and 10000 Hz, have small variations. Notably, at 10000 Hz, the contrast remains almost unity.        3KDVHRIVHFRQGKDUPRQLFRVFLOODWLRQ 2  GHJ +] +] +]   +] Fig. 6. Contrast of interference patterns when the amplitude of a fundamental oscillation is 2 App , where (a) d2 = 0.25, (b) d2 = 0.5, (c) d2 = 0.75, and (d) d2 = 1. To calculate the estimation lines, the measured 2|µn ||b1App (k)|/ℏv values shown in Fig. 5 were used.  D d2  E d2  6.  We presented an approach for analyzing multifrequency sinusoidal oscillating magnetic fields in neutron spin interferometry. We derived a formulation of the resulting interference patterns assuming the oscillating field to be a multifrequency oscillation represented as a Fourier series expansion. The contrast and phase of the interference pattern were expressed as sums of sine functions weighted by Bessel functions. This was confirmed in an experiment with an oscillating magnetic field composed of fundamental and second-harmonic oscillations in the frequency range from 2500 to 10000 Hz. Although the resulting interference patterns are not analytically simple, the experimental data showed reasonable agreement with theoretical predictions, demonstrating the validity of the proposed approach.  3KDVHRILQWHUIHUHQFHSDWWHUQV    F d2  G d2            3KDVHRIVHFRQGKDUPRQLFRVFLOODWLRQ 2  GHJ +] +] +] Conclusions