betaFeedback

WorksPhysics and Astronomy

Crystal and magnetic structures of MnSi in zero magnetic field: Neutron diffraction and symmetry analysis

Abstract

Combining high-resolution wide-angle neutron powder diffraction with magnetic-superspace symmetry analysis, we refine the crystal structure and present, to our knowledge, the first systematic refinement of cubic P213P2_13 MnSi's zero-field magnetic structure across multiple magnetic satellite families. We compare single-arm rhombohedral 3+1D and cubic four-arm 3+4D magnetic superspace-group (MSSG) descriptions with propagation vectors along ⟨111⟩\langle111\rangle. Simultaneous extinction of the two satellite groups around {100}\{100\} dictates a transverse circular-helix model. This suppression and refinements of 18 separately varied satellite-group intensities support this ideal limit with only one adjustable magnetic amplitude. In the ideal single-arm model with MSSG R3.1′(0,0,g)tsR3.1'(0,0,g)ts, the moment has constant magnitude 0.535(3) μBμ_B per Mn. An explicit four-arm 3+4D cubic superspace construction presented here combines four interpenetrating helices and, for the illustrated phase choice, has moments ranging from zero to 0.874(5) μBμ_B per Mn, with the same root-mean-square moment as the 3+1D structure. Their identical powder profiles prevent the present measurements from distinguishing them. The average crystal structure remains cubic within experimental resolution, providing structural motivation for the cubic magnetic model. The mean lattice parameter, determined with formal relative precision σa/a≲10−5σ_a/a \lesssim 10^{-5}, exhibits nonmonotonic temperature dependence below the Néel temperature.

manuscript.pdf

Introduction

MnSi is a prototypical itinerant chiral magnet with the noncentrosymmetric cubic B20 structure, space group P213P2_13. Competition between predominantly ferromagnetic exchange and the Dzyaloshinskii–Moriya interaction favors a long-wavelength modulation, while weak anisotropy selects the ⟨111⟩\langle111\rangle propagation directions [1]. At ambient pressure, magnetic order develops near TN≃29T_N \simeq29 K through a weakly first-order transition [2]. Its field-induced skyrmion lattice has made MnSi a central example of topological magnetism [3], but its zero-field structure nevertheless merits separate examination.

Transport measurements provide complementary evidence for topological spin textures. A topological Hall effect (THE) was reported in the finite-field A phase of MnSi close to TNT_N [4]. In the isostructural short-period helimagnet MnGe, a substantially larger Hall contribution attributed to THE occurs over a broader temperature range [5]. Our zero-field diffraction analysis and the continuum topology discussed below do not establish a connection between the proposed four-arm structure and THE.

Early magnetization and neutron measurements already indicated that MnSi was not a simple ferromagnet [6]. Ishikawa et al. subsequently observed satellites near (110) and (111) using triple-axis diffraction with λ=2.55\lambda= 2.55 Å, and around the reciprocal-space origin using small-angle neutron scattering (SANS) with λ=4.99\lambda= 4.99 Å [7]. They established propagation along ⟨111⟩\langle111\rangle with a period of approximately 180 Å at 4.2 K. Their zero-field diffraction did not, however, determine the spin-rotation plane: a plane perpendicular to the propagation direction was inferred with support from the magnetization. The quoted moment of approximately 0.4μB0.4 \mu_\mathrm{B} per Mn likewise came from magnetization. Thus, the helical interpretation was experimentally well motivated, but was not a refinement of all site-resolved Fourier components against zero-field satellite intensities.

Subsequent polarized small-angle neutron diffraction established the relation between crystal chirality and magnetic helicity, including measurements at 16 K in a 60 Oe guide field after zero-field cooling [8]. These lowfield measurements constrain the helicity of the modulation, but helicity alone does not confirm a single-arm structure. At a resolved magnetic satellite, the polarization dependence probes the Fourier component of the corresponding propagation-vector arm, without determining relative phases between distinct arms. The same helicity signature is therefore compatible with single-arm 3+13+1D domains or with the corresponding arm of a coherent four-arm 3+43+4D structure. Our study addresses these alternatives in zero applied magnetic field.

Polarized SANS further characterized domain selection and the field evolution of the helix [9], while symmetry analyses examined the coupling between structural and magnetic chirality [10]. Complementary, symmetry-constrained muon spin rotation analyses proposed a small relative phase shift between inequivalent Mn orbits of the single-arm structure [11, 12]. These results underline the distinction between establishing the long-period modulation and determining its detailed magnetic structure within the crystallographic unit cell.

To our knowledge, the present work provides the first systematic refinement of the zero-field magnetic structure of MnSi from high-resolution wide-angle neutron powder diffraction intensities across multiple satellite families. Such a refinement requires resolving weak magnetic satellites beside much stronger nuclear reflections and reliably subtracting matched low- and high-temperature patterns. The analysed patterns span approximately Q=0.5Q = 0.5–3.83.8 Å−1^{-1}. The calculated reflection list contains 320 entries grouped into 20 distinct powder positions, providing quantitative intensity constraints on the symmetry-allowed magnetic models.

Our earlier neutron-diffraction and symmetry study of MnGe demonstrated the importance of explicitly comparing symmetry-allowed single- and multi-arm structures [13]. Several models could describe the powder magnetic intensities, whereas the resolved orthorhombic structural distortion supplied a further symmetry constraint. MnSi poses a related question with a different propagation-vector star: four ⟨111⟩\langle111\rangle arms, rather than the three orthogonal ⟨100⟩\langle100\rangle arms considered for MnGe. Here we combine magnetic-superspace symmetry with high-resolution powder diffraction to constrain the zero-field structure of MnSi. We compare a single-arm 3+13+1D description with an explicit four-arm 3+43+4D construction that retains the phase transformations of all four active arms, as detailed in Sec. IV.B. This approach also identifies a remaining limitation: the ideal-helix single-arm and four-arm models supported by the refinements yield identical powder profiles after the appropriate amplitude normalization.

Sample Synthesis and Experimental Methods

The polycrystalline MnSi sample was prepared by arc melting Mn and Si, each of at least 99.99% purity, in a compact MAM-1 arc melter (Edmund Bühler GmbH) [14]. To improve homogeneity, the sample was flipped and remelted several times, then annealed at 1000 °C for 48 h.

Neutron powder diffraction measurements were performed at the SINQ spallation source, Paul Scherrer Institute, Switzerland, using the high-resolution thermal-neutron diffractometer HRPT [15]. Wavelengths of 1.155, 1.494 and approximately 2.45 Å were used. The high-intensity (HI) and high-resolution (HR) modes employ primary white-beam collimations of 40′, and 6′, respectively, with relative resolutions δd/d>1.8×10−3\delta d/d > 1.8 \times10^{-3}, and >0.9×10−3> 0.9 \times10^{-3} [15]. The HR intensity is approximately 7% of the HI intensity. HI mode was used for the weak magnetic diffraction, while HR mode provided greater sensitivity to fine details of the crystal structure.

The magnetic-structure analysis uses three independent HRPT datasets, labelled H1, H2 and H3, measured at λ≃2.45\lambda\simeq2.45 Å in HI mode. Each comprises a base-temperature 1.5–1.8 K pattern and a 54 K reference pattern. H3, used for the representative profile comparison, contains approximately 119 and 90 h of counting at base temperature and 54 K, respectively. The three base-minus-54 K difference patterns were analysed independently on their native grids with 0.05° steps in 2θ2\theta, retaining each dataset’s nuclear scale, angular calibration and peak-profile parameters.

The structural measurements at λ=1.155\lambda= 1.155 Å form a separate HRPT temperature-scan dataset comprising many individual diffraction patterns. For the low-temperature crystal-structure refinement, the datasets collected over 2–6 K were merged to improve counting statistics, yielding the diffraction pattern shown in Figure 1(a). The lattice-parameter curve in Figure 2 is compiled from 110 independently refined patterns, with the individual results and their averages in 1 K intervals shown together.

Figure 1. Nuclear Rietveld refinements of MnSi in P213P2_{1}3: (a) the pattern obtained by merging the 2–6 K datasets measured at HRPT with λ=1.155\lambda= 1.155 Å in HI mode, and (b) the 55 K pattern measured with λ=1.494\lambda= 1.494 Å in HR mode. Black symbols and vertical bars show the data points with standard uncertainties; many bars are smaller than the symbols. Red curves show the calculated profiles, blue dotted curves the observed-minus-calculated differences shifted downwards, and vertical ticks the Bragg positions. The inset in (b) shows the high-angle (440) and (441)/(522) reflections, with the same angular and intensity units as the main panel. The reliability factors are Rwp=4.32%R_{\mathrm{wp}} = 4.32\%, χ2=3.928\chi^{2} = 3.928 for (a), and Rwp=5.53%R_{\mathrm{wp}} = 5.53\%, χ2=3.195\chi^{2} = 3.195 for (b), using FullProf’s conventions.

Figure 2. Temperature dependence of the lattice constant, compiled from individual diffraction patterns at λ=1.155\lambda= 1.155 Å. Grey points show individual refinements; open symbols are averages in fixed 1 K bins. Vertical bars include excess scatter within each bin, and horizontal bars combine the within-bin temperature spread with the temperature variations. The solid lines are guides to the eye. The dashed line marks the literature value TN≃29T_N \simeq29 K. The right-hand axis shows the relative change 100(a−a0)/a0100(a-a_0)/a_0 in percent, with a0=4.5508a_0 = 4.5508 Å.

Complementary measurements were performed on the cold-neutron powder diffractometer DMC at λ≃2.45\lambda\simeq2.45 Å. A base-minus-54 K dataset, on its native 0.125° grid in 2θ2\theta, provides a high-statistics check of the absence of the first {100}\{100\} satellites. DMC was used for the magnetic temperature scans because its full-detector count rates estimated from MnSi nuclear peaks were approximately 50–60 times those of HRPT in HI mode, allowing useful statistics in much shorter reasonable counting intervals. The quantitative magnetic profile fits are based on HRPT datasets. Experimental details, DMC profile diagnostics and subtraction limitations, and crystal-structure checks are provided in § 13–§ 17 of the Supplementary Material (SM) [16]. Hereafter, S-prefixed section, figure and table numbers refer to the SM.

FullProf [17] was used for crystal- and magnetic-structure refinements and for calculating diffraction profiles, using its tabulated neutron scattering lengths. The complementary weighted profile-decomposition method is described in Sec. IV.A. Symmetry analysis was performed using ISODISTORT from the ISOTROPY software suite. ware suite [18, 19] and tools of the Bilbao Crystallographic Server, including MVISUALIZE [20, 21].

Crystal Structure

The nuclear diffraction patterns are described by the cubic space group P213P2_{1}3 (No. 198), with Mn and Si at 4a positions (x,x,x)(x, x, x). Figure 1 compares the merged 2–6 K pattern measured at λ=1.155\lambda= 1.155 Å in HI mode with the 55 K pattern measured at λ=1.494\lambda= 1.494 Å in HR mode. With both sites fully occupied, the short-wavelength refinement gives xMn=0.86358(9)x_{\mathrm{Mn}} = 0.86358(9) and xSi=0.15454(7)x_{\mathrm{Si}} = 0.15454(7), isotropic displacement parameters BMn=0.103(9)B_{\mathrm{Mn}} = 0.103(9) and BSi=0.094(8)B_{\mathrm{Si}} = 0.094(8) Å2^{2}, and a=4.55109(7)a = 4.55109(7) Å.

The HR patterns at 1.8 and 55 K show no resolved peak splitting. At high angles, the peaks are as narrow as those described by the Na2_{2}Al2_{2}Ca3_{3}F14_{14} (NAC) resolution calibration standard [22, 23], with no positive strain broadening required. The refined corrections to the Gaussian tan⁡2θ\tan^{2}\theta coefficient in the Caglioti width function [24] are U=−0.00382(35)U = -0.00382(35) and −0.00401(34)-0.00401(34) deg2^{2} at 1.8 and 55 K, respectively, both measured in HR mode. These are empirical corrections to the instrumental profile, with VV and WW also refined, rather than physical negative strains. The mean HR lattice parameter is determined with a formal relative precision better than 10−510^{-5} with σa/a≃5.5×10−6\sigma_{a}/a \simeq5.5 \times10^{-6}. This uncertainty and the small scatter of the HI temperature-scan values quantify the precision of the mean cubic cell (Figure 2).

Selecting a single propagation-vector arm along ⟨111⟩\langle111 \rangle lowers the magnetic symmetry to trigonal and, through magnetoelastic coupling, is generally expected to induce a rhombohedral distortion. No such distortion is detected in our HR data. The retained cubic metric therefore provides a structural motivation for the cubic magnetic superspace model. Distinguishing it conclusively from a single-arm state requires comparing the theoretically predicted magnetostrictive distortion with a quantitative experimental detection limit. For comparison, our study of isostructural MnGe resolved an orthorhombic structure, P212121P2_{1}2_{1}2_{1} (No. 19), in the magnetically ordered state below TN≃170T_{N} \simeq170 K [13]. At 2 K, the relative spread of the lattice parameters is (c−a)/aˉ≃1.6×10−3(c-a)/\bar{a} \simeq1.6 \times10^{-3}, where aˉ=(a+b+c)/3\bar{a} = (a+b+c)/3. The resolved distortion favoured orthorhombic magnetic superspace descriptions over the cubic model. No analogous splitting of this size is resolved in MnSi. However, even assuming similar magnetoelastic coupling strengths, the induced strain also depends on the elastic stiffnesses. and the magnitude and symmetry of the magnetic order parameter. The MnGe result therefore illustrates the value of structural information for distinguishing magnetic models, but does not provide a quantitative prediction of the rhombohedral distortion in a single-arm MnSi state.

Site occupancies were assessed using the λ=1.155\lambda= 1.155 Å pattern, which extends to Q≃10.7Q \simeq10.7 Å−1^{-1}. This wide QQ range helps separate occupancy effects from those of atomic displacement parameters (ADPs): occupancy scales the atomic scattering contribution, whereas ADPs progressively attenuate it with increasing QQ. With the Si site kept fully occupied, the Mn occupancy refines to 0.965(14). Allowing this parameter to vary gives only a small improvement: χ2\chi^2 decreases from 3.9281 to 3.9247 and RwpR_{\mathrm{wp}} from 4.32% to 4.31%. Correlated profile residuals persist, and systematic errors may affect the refined occupancy; the apparent Mn deficit therefore does not establish the presence of vacancies. We describe the sample as near-stoichiometric and retain fully occupied sites as the reference model in Fig. 1; the refinements do not prove exact 1:1 stoichiometry. This qualification is relevant for MnSi: studies of controlled starting compositions identify Mn vacancies and antisite disorder, with changes in the magnetic transition, even when bulk composition measurements appear stoichiometric within their uncertainty [25, 26].

The lattice constant has a broad maximum near 10 K (Figure 2), decreases on warming towards the magnetic transition, and has a shallow minimum in the 30–40 K range before increasing again. The contraction on warming towards TNT_N, followed by expansion above the transition region, is qualitatively consistent with the thermal-expansion measurements of Matsunaga et al. [27]. Their results do not, however, establish the broad low-temperature maximum observed in our diffraction scan. Dilatometry resolves a small discontinuity near 29 K superposed on a broader thermal-expansion anomaly [2, 28]. Our diffraction scan is consistent with a nonmonotonic lattice response, but does not resolve this small discontinuity or establish a separate transition near 10 K. Small variations of instrumental wavelength can also affect changes of this size and are not included in the pointwise refinement errors.

The measured a(T)a(T) provides a practical guide for selecting reference temperatures for magnetic subtraction. In particular, 54 K is above TNT_N and has a lattice constant close to that near base temperature, reducing shifts of the nuclear peaks between the two patterns. Matching the lattice constants improves cancellation; it does not guarantee identical intensities or line shapes.

Magnetic structure determination and discussion

The magnetic satellites in MnSi lie close to much stronger nuclear Bragg reflections. We therefore use the difference patterns between the base-temperature and 54 K diffraction patterns measured on HRPT for the magnetic structure determination. This subtraction suppresses the nuclear contribution, although imperfect cancellation can leave residual features that must be considered when assessing the magnetic fits. The subtraction also suppresses scattering from minor impurity phases, provided that their diffraction contribution remains approximately unchanged between base temperature and 54 K.

The magnetic signal requires percent-level discrimination against the nuclear scattering. Across H1–H3, the complete {110}\{110\}, {111}\{111\} and {210}\{210\} satellite groups have integrated areas of approximately 1.6%, 1.1–1.4% and 1.9–2.1%, respectively, of the corresponding 54 K nuclear peak areas. The ratio is larger, about 8%, for the weaker nuclear {211}\{211\} reflection. For a magnetic contribution of 1–2%, measuring its area to 10% precision requires residual nuclear contributions to be controlled at roughly 0.1–0.2% of the nuclear area. This motivates long counting times, matched-temperature references and comparison of independent measurements.

The three HRPT datasets provide independent tests of the magnetic intensity distribution. No additional peak near 51.5∘51.5^{\circ} or negative-flank features are resolved in H1–H3. Such features, observed in some measurements with both HRPT and DMC, are considered separately in Sec. VI and are not assigned to intrinsic MnSi magnetic satellites.

Le Bail fits and weighted profile decomposition

We first tested the indexing of the magnetic difference patterns by le Bail profile fitting in FULLPROF [17]. In this procedure the integrated reflection intensities are extracted without imposing a magnetic structure-factor model. The satellite positions are specified in the parent cubic reciprocal basis as H±k\mathbf{H} \pm\mathbf{k}, where H\mathbf{H} is a reciprocal-lattice vector of the parent primitive cubic lattice and k\mathbf{k} belongs to the star of k1=[k,k,k]\mathbf{k}_1 = [k,k,k] along the cubic ⟨111⟩\langle111\rangle directions (the Λ\Lambda line, labelled LD). We follow the established nomenclature for magnetic irreducible representations (irreps) and magnetic superspace groups (MSSGs) [18, 20].

The absence of resolved magnetic peaks at the first two satellite groups around {100}\{100\}, expected near 2θ≃30.8∘2\theta\simeq30.8^{\circ} and 31.7∘31.7^{\circ}, provides a very strong constraint. The DMC data offer a particularly sensitive check in this region [in-set of Figure 3(a)]. Neither group is resolved above the local background within experimental sensitivity. Writing I100I_{100} for the summed area of both first satellite groups, a local weighted fit gives I100/I110=0.0064(39)I_{100}/I_{110} = 0.0064(39), consistent with zero. Here I110I_{110} is the complete {110}\{110\} satellite area.

Figure 3. HRPT H3 base-minus-54 K profile: (a) full fitted interval; (b,c) enlarged {110}\{110\} and {111}\{111\} satellite groups. Points show measured intensities with counting-error bars; blue and dashed red curves are the weighted free-intensity and ideal-helix fits. The same ideal-fit background is removed from all curves and intensities normalized by the scale factor refined from the 54 K nuclear pattern. The ideal residual is offset in (a). The nominal goodness-of-fit values are χ2=1.3590\chi^2 = 1.3590 and 1.3887, respectively. The inset shows the DMC difference near the absent {100}±k\{100\} \pm\mathbf{k} groups, with its fixed background removed; vertical dotted lines mark their expected positions. The inset retains DMC intensity units and its original counting errors. Short ticks in the main panels mark the indexed satellite positions at the fitted H3 k\mathbf{k}.

Le Bail extraction updates intensities by iterative partitioning [17]; for weak reflections, counting noise and the allocation of intensity between overlapping peaks and background can leave small positive areas without a resolved peak. We therefore use direct weighted whole-pattern decomposition, with satellite intensities and background optimized simultaneously, as the quantitative reference for comparing magnetic models (Sec. S4 A).

FullProf supplied the calculated satellite profiles, with lattice, angular and profile parameters fixed to each dataset’s 54 K nuclear refinement and kk fixed to its ideal-helix estimate: 0.01433(19), 0.01520(20) and 0.01467(12) for H1, H2 and H3, respectively, in reciprocal-lattice units. After imposing the two {100}\{100\} extinctions, 296 reflection entries form 18 groups of coincident reflections: within each group, the reflections have the same calculated dd-spacing and hence the same 2θ2\theta for the cubic lattice and chosen propagation-vector star. One fitted coefficient describes their combined intensity. Writing Pg(2θ)P_g(2\theta) for the fixed calculated profile of group gg, the fitted intensity is Iicalc=∑g=118cgPg(2θi)+b(2θi)I_i^{\mathrm{calc}} = \sum_{g=1}^{18} c_g P_g(2\theta_i) + b(2\theta_i). Here ii runs over all measured angular bins, whereas gg labels the groups. The function b(2θ)b(2\theta) is a single quadratic background across the entire fitted range, with the same functional form as in the ideal-helix FullProf fit. The 18 group coefficients and three background coefficients were optimized simultaneously by weighted least squares. The measured points around the two extinct {100}\{100\} groups remained included in the fit. Figure 3 compares this decomposition with the ideal helix for H3, the dataset with the smallest statistical uncertainties.

The free-intensity decomposition provides a reference for the best fit quality attainable under the adopted calibration, profile, background and extinction constraints, without imposing additional magnetic intensity relations between the retained groups.

The DMC le Bail refinement gives the propagation-vector component k=0.01510(10)k = 0.01510(10).

Symmetry analysis and magnetic models

The parent space group P213P2_13 (No. 198) has three magnetic irreducible representations (irreps) associated with the LD line (k,k,k)(k,k,k). The pure mLD1 model fits the measured magnetic profiles poorly and has an independent difficulty: threefold symmetry forces its structure factor at Q=k1\mathbf{Q} = \mathbf{k}_1 to be longitudinal. Its perpendicular component therefore vanishes, in conflict with the observed small-angle satellites [7, 9]. We consequently consider mLD2 and its complex-conjugate partner mLD3, corresponding to opposite helicities that cannot be distinguished by these unpolarized powder measurements.

The propagation-vector star consists of four symmetry-equivalent wave vectors k1=(k,k,k)\mathbf{k}_1 = (k,k,k), k2=(k,−k,−k)\mathbf{k}_2 = (k,-k,-k), k3=(−k,k,−k)\mathbf{k}_3 = (-k,k,-k) and k4=(−k,−k,k)\mathbf{k}_4 = (-k,-k,k), which satisfy the linear relation ∑iki=0\sum_i \mathbf{k}_i = 0. Consequently, only three propagation vectors are linearly independent, although all four arms are required to preserve the full cubic symmetry of the propagation-vector star. For the selected mLD2 order parameter, however, the conventional 3+3D MSSG generated by ISODISTORT at a fixed phase sum is orthorhombic, as explained below.

For a single active arm k1\mathbf{k}_1, the mLD2 model has rhombohedral MSSG R3.1′(0,0,g)tsR3.1'(0,0,g)_{\mathrm{ts}} (No. 146.1.22.2.m11.2). In its standard RR-centred hexagonal setting, the basis is

ah=bc−cc,bh=ac−bc,\mathbf{a}_h=\mathbf{b}_c-\mathbf{c}_c,\qquad\mathbf{b}_h=\mathbf{a}_c-\mathbf{b}_c,
ch=−ac−bc−cc,\mathbf{c}_h=-\mathbf{a}_c-\mathbf{b}_c-\mathbf{c}_c,
kh=(0,0,g),g=−3k,(1)\mathbf{k}_h=(0,0,g),\qquad g=-3k, \tag*{(1)}

with zero origin shift. The subscripts c,hc,h denote parent cubic and hexagonal axes; ah=2ac\mathbf{a}_h=\sqrt{2}\mathbf{a}_c, ch=3ac\mathbf{c}_h=\sqrt{3}\mathbf{a}_c and γh=120∘\gamma_h=120^\circ. The Mn 4a orbit splits into 3a and 9b, allowing two and six independent real first-harmonic coefficients, respectively; fixing the unobservable common phase leaves seven magnetic variables.

The mLD1 exclusion, rhombohedral conversion conventions and matched R3R3/cubic refinements are detailed in Sec. S5.

Following a suggestion by B. J. Campbell [29], the order-parameter direction (a,b;a,b;a,b;a,b)(a,b;a,b;a,b;a,b) for irrep mLD2LE2 remains invariant under the threefold rotation about the cubic [111][111] axis only when the rotational part is combined with the superspace phase translation (−1/3,−1/3,−1/3,−1/3)(-1/3,-1/3,-1/3,-1/3). Importantly, these superspace translations are associated with the four active propagation vectors rather than with the three independent modulation coordinates.

Using this observation, we construct the cubic 3+4D superspace operators from the generators of the parent space group and the representation matrices of the 8-dimensional mLD2LE2 irrep. Each matrix was applied to the order-parameter direction (a,b;a,b;a,b;a,b)(a,b;a,b;a,b;a,b), and compensating translations of the four internal coordinates were introduced where required to restore invariance. For the threefold rotation about [111][111], the phase factor exp⁡(2πi/3)\exp(2\pi i/3) is compensated by the internal translation (2/3,2/3,2/3,2/3)(2/3,2/3,2/3,2/3), equivalent modulo one to the translation given above. This procedure yields the operators in Table 1.

References

  1. [1]P. Bak and M. H. Jensen, Journal of Physics C: Solid State Physics 13, L881 (1980).
  2. [2]S. M. Stishov, A. E. Petrova, S. Khasanov, G. K. Panova, A. A. Shikov, J. C. Lashley, D. Wu, and T. A. Lograsso, Physical Review B 76, 052405 (2007).
  3. [3]S. Mühlbauer, B. Binz, F. Jonietz, C. Pfleiderer, A. Rosch, A. Neubauer, R. Georgii, and P. Böni, Science 323, 915 (2009), https://www.science.org/doi/pdf/10.1126/science.1166767, URL https://www.science.org/doi/abs/10.1126/science.1166767.
  4. [4]A. Neubauer, C. Pfleiderer, B. Binz, A. Rosch, R. Ritz, P. G. Niklowitz, and P. Böni, Phys. Rev. Lett. 102, 186602 (2009), URL https://doi.org/10.1103/PhysRevLett.102.186602.
  5. [5]N. Kanazawa, Y. Onose, T. Arima, D. Okuyama, K. Ohoyama, S. Wakimoto, K. Kakurai, S. Ishiwata, and Y. Tokura, PHYSICAL REVIEW LETTERS 106, 156603 (2011), ISSN 0031-9007.
  6. [6]L. M. Levinson, G. H. Lander, and M. O. Steinitz, AIP Conference Proceedings 10, 1138 (1973).
  7. [7]Y. Ishikawa, K. Tajima, D. Bloch, and M. Roth, Solid State Communications 19, 525 (1976).
  8. [8]M. Ishida, Y. Endoh, S. Mitsuda, Y. Ishikawa, and M. Tanaka, Journal of the Physical Society of Japan 54, 2975 (1985).
  9. [9]S. V. Grigoriev, S. V. Maleyev, A. I. Okorokov, Y. O. Chetverikov, P. Böni, R. Georgii, D. Lamago, H. Eckerlebe, and K. Pranzas, Physical Review B 74, 214414 (2006).
  10. [10]V. Dmitriev, D. Chernyshov, S. Grigoriev, and V. Dyadkin, Journal of Physics: Condensed Matter 24, 366005 (2012).
  11. [11]P. Dalmas de Réotier, A. Maisuradze, A. Yaouanc, B. Roessli, A. Amato, D. Andreica, and G. Lapertot, Physical Review B 93, 144419 (2016).
  12. [12]P. Dalmas de Réotier, A. Yaouanc, A. Amato, A. Maisuradze, D. Andreica, B. Roessli, T. Goko, R. Scheuermann, and G. Lapertot, Quantum Beam Science 2, 19 (2018).
  13. [13]V. Pomjakushin, I. Plokhikh, J. S. White, Y. Fujishiro, N. Kanazawa, Y. Tokura, and E. Pomjakushina, Physical Review B 107, 024410 (2023).
  14. [14]Edmund Bühler GmbH, Compact arc melter MAM-1, https://www.edmund-buehler.de/en/materials-science/arc-melting/compact-arc-melter-mam-1, accessed 14 October 2026.
  15. [15]P. Fischer, G. Frey, M. Koch, M. Koennecke, V. Pomjakushin, J. Schefer, R. Thut, N. Schlumpf, R. Buerge, U. Greuter, et al., Physica B 276-278, 146; https://www.psi.ch/sinq/hrpt (2000).
  16. [16]See Supplemental Material appended to this manuscript for experimental datasets, profile and structural checks, refinement details, temperature-dependent intensities, and superspace and continuous-field derivations.
  17. [17]J. Rodriguez-Carvajal, Physica B 192, 55; www.ill.eu/sites/fullprof/ (1993).
  18. [18]B. J. Campbell, H. T. Stokes, D. E. Tanner, and D. M. Hatch, J. Appl. Cryst. 39, 607 (2006).
  19. [19]H. T. Stokes, D. M. Hatch, and B. J. Campbell, ISOTROPY Software Suite, iso.byu.edu (2021).
  20. [20]M. Aroyo, J. Perez-Mato, D. Orobengoa, E. Tasci, G. De la Flor, and A. Kirov, Bulgarian Chemical Communications 43, 183; Bilbao Crystallographic Server http://www.cryst.ehu.es (2011).
  21. [21]J. M. Pérez-Mato, S. V. Gallego, E. S. Tasci, L. Elcoro, G. de la Flor, and M. I. Aroyo, Annual Review of Materials Research 45, 217 (2015).
  22. [22]F. Gozzo, L. De Caro, C. Giannini, A. Guagliardi, B. Schmitt, and A. Prodi, J. Appl. Cryst. 39, 347 (2006).
  23. [23]Paul Scherrer Institute, HRPT data refinements, https://www.psi.ch/en/sinq/hrpt/data-analysis, accessed 2 November 2026.
  24. [24]A. W. Hewat, Nuclear Instruments and Methods 127, 361 (1975).
  25. [25]M. Reiner, A. Bauer, M. Leitner, T. Gigl, W. Anwand, M. Butterling, A. Wagner, P. Kudejova, C. Pfleiderer, and C. Hugenschmidt, Scientific Reports 6, 29109 (2016).
  26. [26]T. Y. Ou-Yang, G. J. Shu, J.-Y. Lin, C. D. Hu, and F. C. Chou, Journal of Physics: Condensed Matter 28, 026004 (2016).
  27. [27]M. Matsunaga, Y. Ishikawa, and T. Nakajima, Journal of the Physical Society of Japan 51, 1153 (1982).
  28. [28]A. E. Petrova and S. M. Stishov, Physical Review B 94, 020410 (2016).
  29. [29]B. J. Campbell (2026), private communication.
  30. [30]H. T. Stokes and D. M. Hatch, Isotropy Subgroups of the 230 Crystallographic Space Groups (World Scientific, 1988).
  31. [31]S. Kirkpatrick, C. D. Gelatt, and M. P. Vecchi, Science 220, 671 (1983).
  32. [32]K. Shimizu, S. Okumura, Y. Kato, and Y. Motome, Phys. Rev. B 105, 224405 (2022), URL https://doi.org/10.1103/PhysRevB.105.224405.
  33. [33]V. A. Chizhikov and V. E. Dmitrienko, Phys. Rev. B 85, 014421 (2012).
  34. [34]V. A. Chizhikov and V. E. Dmitrienko, Phys. Rev. B 88, 214402 (2013).
  35. [35]K. P. W. Hall and S. H. Curnoe, Phys. Rev. B 104, 094408 (2021).
  36. [36]M. L. Plumer and M. B. Walker, J. Phys. C: Solid State Phys. 15, 7181 (1982).
  37. [37]M. A. Wilde, M. Dodenhöft, A. Niedermayer, A. Bauer, M. M. Hirschmann, K. Alpin, A. P. Schnyder, and C. Pfleiderer, Nature 594, 374 (2021).
  38. [38]T. Adams, S. Mühlbauer, C. Pfleiderer, F. Jonietz, A. Bauer, A. Neubauer, R. Georgii, P. Böni, U. Keiderling, K. Everschor, et al., Phys. Rev. Lett. 107, 217206 (2011).
  39. [39]J. Kindervater, I. Stasinopoulos, A. Bauer, F. X. Haslbeck, F. Rucker, A. Chacon, S. Mühlbauer, C. Franz, M. Garst, D. Grundler, et al., Phys. Rev. X 9, 041059 (2019).

Paper details

vlnc:W_pjcbu8rj
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents