WorksPhysics and Astronomy
Magnetism and magnetoelastic effect in 2D van der Waals multiferroic CuCrP2S6
Abstract
We report a magnetic and neutron diffraction study on the ground state magnetism and field evolution of single crystal van der Waals multiferroic CuCrP2S6. The ordered moments align along the b axis in the A-type antiferromagnetic configuration with a spin-flop transition along the same direction. Field application along a introduces a smooth transition to a fully-polarized ferromagnetic state via in-plane spin rotation. These findings resolve the ambiguity of the ground state magnetization direction in CuCrP2S6 and uncover its field responses, providing a firm basis for future magnetoelectric study. A magnetoelastic coupling effect connecting the interlayer spacing and the magnetic order was further revealed, highlighting the out-of-plane strain as an effective control knob for tuning magnetism both in this system and in related van der Waals magnets.
Introduction
Ever since the discovery of intrinsic two-dimensional (2D) magnetism in monolayer FePS [1, 2], and CrI [3], the exploration of magnetism in van der Waals (vdW) layered materials has become a central focus in advancing spintronic technologies [4, 5]. Within this context, metal thio/selenophosphates with a general formula MPX (M = Fe, Mn, Ni, X = S, Se) have attracted extensive attention as a versatile material platform, owing to its chemically tunable magnetic interactions [6] [7]. Introducing heterometallic cations enriches this material family with polar, electronic, and charge degrees of freedom in addition to magnetism, giving rise to diverse functional behaviors including ferroelectric switching [8], optoelectronic response [9, 10], anisotropic electrical conductance [11, 12]. The interplay among these degrees of freedom further promotes multifunctionality [13, 14]. CuCrPS (CCPS) is a particularly notable example, exhibiting coexisting antiferroelectricity (AFE) and antiferromagnetism (AFM) with remarkable polarization-magnetization coupling [11, 15–19].
CCPS crystallizes in the monoclinic structure (space group ) at ambient condition [20], with lamellar layers of edge-sharing CrS, PS and CuS octahedra that form interconnected triangular lattices Figure 1(a). Cr () locates at the octahedra center, carrying an effective magnetic moment of Bohr magneton (). Early magnetization and powder neutron diffraction studies identified an “A-type” AFM order below K, consisting of ferromagnetically aligned basal layers stacked antiferromagnetically along the layer normal direction [21, 22]. In contrast, the nonmagnetic Cu resides in an effective double-well potential [20, 21, 23] arising from the pseudo Jahn–Teller effect [23–26]. This bistable configuration of CuS gives rise to electric-field-tunable polarization [27, 28], offering potential for memory devices. More importantly, the magnetoelectric coupling (MEC) effect between the polar order (Cu) and magnetic order (Cr) provides a natural setting for multiferroicity, which has prompted extensive theoretical and experimental investigations [16, 18, 19].

Figure 1. (a) Crystal structure of CuCrPS in the AFE state. (b) Field cool magnetization and along the and crystallographic axes, respectively. Inset shows the refined magnetic structure. (c) Field dependence of at selective temperatures. Inset highlights the spin-flop transition in the low field regime. (d) Derivative of the with respect to the field strength. The peak position defines the spin-flop transition critical field .
Recently, we elucidated the microscopic ordering mechanism of the Cu sublattice in CCPS [29], thereby establishing a foundation for investigating MEC in this material. Nevertheless, a definitive characterization of the magnetic ground state remains elusive. Early work suggested the ordered moments are oriented along the in-plane diagonal direction [21, 22], whereas more recent reports present conflicting assignments, placing the ordered Cr spins along the [11, 17], the axes [30], or leaving the moment orientation unspecified [14, 16, 18, 31]. Furthermore, the direction-dependent evolution of the magnetic structure under applied field remains unresolved. This lack of consensus on the magnetic ground state, together with the absence of studies on its field response, presents a fundamental barrier to understanding both the intrinsic magnetism of CCPS and its proposed multiferroic behavior.
In this article, we present a comprehensive neutron diffraction and magnetization study of the magnetic ground state of CCPS and its field response. We identify a low temperature “A-type” AFM spin configuration with magnetic easy axis along the axis. A spin-flop transition occurs under field applied along the same direction. In contrast, magnetic field applied along drives a continuous spin rotation towards a fully-polarized ferromagnetic (FM) state. A magnetoelastic coupling between the spin canting angle and the interlayer spacing is also revealed, suggesting a new route for controlling magnetism through out-of-plane strain in this vdW antiferromagnet.
Experimental details
Single crystals CCPS were synthesized from pure elements following the procedure described in Ref. [32]. Magnetic measurements were performed using a Quantum Design magnetic property measurement system (MPMS) on a mg crystal, with the crystallographic directions determined by x-ray Laue diffraction. Single crystal x-ray measurements were carried out on a CCPS flake with dimensions of mounted to a Mitegen loop using Paratone-N oil. Data were collected on a Rigaku Synergy-DW diffractometer which uses Mo source with wavelength of 0.7107 Å. The sample temperature is regulated using Oxford cryostream. The reduced data were processed with the CrysAlisPro software [33]. Neutron diffraction measurements were performed on the time-of-flight single crystal diffractometer CORELLI at the Spallation Neutron Source (SNS) at Oak Ridge National Laboratory (ORNL) [34]. 22 pieces of single crystals (mass mg) were co-aligned on an aluminum plate mounted to a sample stick. The sample mosaic was estimated to be less than . For the measurements in zero field, a closed-cycle refrigerator was used with a base temperature of K. For field-dependent measurements, a 5-Tesla superconducting cryomagnet was used. The reciprocal space was surveyed by rotating the sample stick at steps at each temperature and field. The temperature and field dependence of the magnetism was characterized via monitoring the magnetic peak intensity at constant ramping rate.
Results
Figure 1(b) presents the temperature dependence of and near the magnetic transition. Both increase as the temperature approaches from above. At K, exhibits a sharp -like cusp, whereas maintains nearly constant below . These features are characteristic of an AFM transition with moments along the direction [35], consistent with Ref. [30]. Figure 1(c) shows the field dependence of at various temperatures. At K, displays a low-field anomaly, followed by a linear trend and eventual saturation at T. Such anomalies are commonly associated with a spin-flop transition [17, 30, 36], in which a magnetic field applied along the Néel vector induces its rotation to minimize the Zeeman energy. This observation further supports that the ordered moments are aligned along the direction in the ground state. As approaches , the saturation field decreases gradually, and the spin-flop transition becomes increasingly broadened.
The critical field , defined as the peak position in the derivative of with respect to (), is determined to be 0.44 T at K [Fig. 1](d), in good agreement with Ref. [30]. It decreases to T and 0.35 T at K and 20 K, respectively, and is expected to approach zero at as the magnetic order vanishes. The saturation field also decreases with increasing temperature, reflecting a progressive weakening of the exchange interactions due to thermal fluctuations. Because the intralayer FM interaction is dominant in CCPS [21, 30], its magnetic structure can be approximated as a stacking of rigid ferromagnetic layers coupled antiferromagnetically within a mean-field framework. This description enables extraction of the effective anisotropy field and exchange field using the relations and [37].
These parameters characterize the magnetic anisotropy within individual layers and the interlayer exchange stiffness. They are related to the single ion anisotropy constant and interlayer AFM exchange coupling through and . The extracted values are summarized in Table I, which enables modeling of [38] and determination of the field dependence of the canting angle between the adjacent FM layers, to be discussed in Figure 3(b).
| (K) | (T) | (T) | (/Cr) |
| 2 | 3.36 | 0.03 | 2.90 |
| 10 | 3.02 | 0.03 | 2.75 |
| 20 | 2.51 | 0.02 | 2.33 |
Table I. Extracted exchange field , anisotropy field , and saturation magnetic moments from the magnetization data at different temperatures based on the stacking model with AFM-coupled rigid FM layers.

Figure 3. Field evolution of the magnetic structure of CCPS with -axis in-plane field. (a) Field dependence of the and peak intensities at K. Insets illustrate the evolution of the magnetic structure with increasing field. The solid line is the model fit and the dashed line is guide to the eye. (b) Field dependence of the in-plane spin canting angle . Symbols are experimental values obtained via fitting the peak intensity at different temperatures, solid lines are model predicted values using the stacking model. The dashed line marks the fully-polarized state with .
Neutron diffraction further confirms the direction of the ordered magnetic moment. Figure 2(a) presents the diffraction patterns in the plane at 5 and 50 K. In addition to the nuclear Bragg peaks at integer positions, weak fractional peaks appear at and , arising from twinned structural domains in the pseudo-hexagonal symmetry in the basal plane [38]. At K, strong magnetic reflection appear at positions indicating AFM coupling between adjacent layers. Fractional magnetic peaks originating from other structural domains are also present (marked by arrows). The temperature dependence of the peak intensity yields K, in good agreement with the magnetization data [Figure 2(b)]. The extracted component suggests a 3D critical behavior near the transition. Together with the absence of magnetic diffuse scattering (indicating negligible critical fluctuation), the presence of significant interlayer magnetic interactions, and weak in-plane anisotropy, demonstrate that CCPS is an anisotropic 3D magnetic system.

Figure 2. (a) Neutron diffraction pattern in the scattering plane at K (top) and K (bottom) at zero field. Orange arrows mark the magnetic peaks from other structural domains. (b) The temperature dependence of the magnetic peak intensity at zero field. The solid line is a power law fit of with the reduced temperature . The upper inset shows the co-aligned single crystal array used in the neutron measurements. The lower inset shows the log − log plot with a linear fit in the range to determine .
Detailed magnetic structure refinement based on representational analysis reveals an “A-type” AFM configuration (magnetic space group ), consisting of FM layers within the plane that stack antiferromagnetically along the layer normal [[38]]. The ordered moments of 3.03(5) , align along the direction, in agreement with the magnetization results [Figure 1(b)]. Notably, the absence of magnetic scattering at position provides additional support for this moment orientation, since the magnetic neutron scattering is sensitive only to the spin component perpendicular to the momentum transfer, .
To characterize the field evolution of magnetism in CCPS, neutron diffraction measurements were carried out under a magnetic field applied along the direction (w.r.t. the main structural domain). Figure 3(a) shows the field dependence of the and peak intensities at K. As the field increases, the intensity of the peak decreases progressively and vanishes above a critical field T. Concurrently, the peak emerges and gradually saturates near the same field. The zero-field intensity is fully recovered upon removing the field, implying negligible hysteresis. Upon cooling, increases to approximately 4.5 T at K and quickly exceeds the accessible field range [[38]]. This change in magnetic intensity can be understood as a continuous, symmetry-adapted evolution from the AFM configuration to a field-polarized FM state with moments aligned along the [Figure 3(a) inset].
To quantify this evolution, we adopt the same simplified stacking model used in the analysis of the magnetization data and employed an empirical relation, , to describe the canting angle between the FM moment and the direction as a function of magnetic field , where , , are constants [[38]]. By calculating the magnetic structure factor at a given , the magnetic peak intensity can be fitted, allowing to be extracted as a function of . Figure 3(b) plots the resulting values at selected temperatures and fields. Below K, remains below up to T, indicating robust interlayer AFM coupling against thermal fluctuation. As the temperature increases, the field response of becomes progressively softer. Full polarization is reached at T at K, consistent with the disappearance of the peak intensity and the saturated peak.
Energetic analysis within the stacking model shows that the canting angle can be determined from the relation for a magnetic field applied along the direction [[38]]. Using the parameters listed in Table I, the calculated field dependence of agrees well with the values obtained from fitting the neutron data [Figure 3(b)]. Applying this expression and assuming negligible anisotropy at K yields an estimate of T, highlighting the rapid evolution of exchange coupling as the system approaches .
Figure 4(a) presents the -dependence of the magnetic peak intensity at T along . Compared to the zero-field case, the AFM transition is suppressed to K. Although a net moment along is symmetry-allowed, the reduced implies that this configuration is energetically less favorable, consistent with the magnetic easy axis lying along the direction. The onset of the AFM order is accompanied by a broad maximum in the interlayer spacing between adjacent planes. The increase in interlayer spacing upon cooling reflects negative thermal expansion (NTE) along the layer normal direction [32]. In contrast, the rapid decrease below indicates an energy minimization driven by AFM ordering. The observed hump in the interlayer spacing therefore arises from the combined effect of NTE and magnetic ordering. Below , applying field along leads to an expansion of the interlayer spacing Figure 4(b). This behavior can be understood as a competition between the interlayer AFM coupling and the Zeeman energy; increasing field enhances the canting angle, thereby increasing the net FM moment and imposing an energy penalty on . As a result, the interlayer spacing expands to minimize the total energy.

Figure 4. Magnetoelastic coupling effect. (a) Temperature dependence of (0, 0, 1) magnetic peak intensity and the interlayer spacing under T along the direction. The solid line is a power law fit. (b) Field evolution of the interlayer spacing at K and 22 K.
Based on the neutron scattering and magnetization results, Figure 5 summarizes the magnetic phase diagram of CCPS. At high temperature, the sublattice is paramagnetic. Below K, an “A-type” AFM phase is stabilized, with the ordered moment aligned along . When a magnetic field is applied along , exceeding a critical field induces a spin-flop transition, in which the staggered moments rotate toward while retaining a net component along . In contrast, a field along the direction drives a symmetry-adapted canted AFM phase with a net moment along . In both field orientations, a sufficiently strong field fully polarizes the sublattice.

Figure 5. - phase diagram of the magnetic structure in CCPS. Insets illustrate the magnetic structure at T, the spin-flop AFM configuration at . Colored symbols are obtained from neutron diffraction results with field applied along the direction.
Discussions
The pseudo-hexagonal monoclinic metric enables the formation of rotated structural domains, a phenomenon well-known in layered chalcogenides and halides, where twisted stacking gives rise to fractional diffraction peaks [39]. This effect highlights the difficulty of obtaining reliable single crystal magnetization data and may account for the lack of consensus regarding the ordered moment direction in recent literature. The misassignment of ordered moment direction in earlier powder neutron studies likely originates from peak overlap between (0, 2, 0) and (-1, 1, 0) reflections, caused by their close scattering angles and the limited instrument resolution [22].
The observed magnetic anisotropy is relatively weak, as evidenced by the small value of . In general, magnetic anisotropy arises from three primary contributions: shape anisotropy, single-ion anisotropy, and exchange anisotropy [40]. Although shape anisotropy can be significant in 2D materials and may favor in-plane magnetization in CCPS, it cannot account for the easy-axis alignment along , particularly given the irregular sample geometry used in this work. Single-ion anisotropy arises from a combined effect of spin-orbit coupling (SOC) and crystal electric field (CEF). In an approximate trigonal-distorted octahedron environment, adopts a half-filled low energy manifold in the CEF ground state, similar to that in and [41]. The absence of orbital degeneracy leads to a quenched orbital moment [42, 43], as evidenced by the nearly isotropic Landé factor [21, 30] and an ordered moment , in excellent agreement with the classical spin-only limit. This indicates that single ion anisotropy originates from second order SOC via quantum fluctuations of the orbital moment [42]. Consequently, the anisotropy is expected to be weak, scaling as , where is the SOC strength for and denotes the energy separation between the CEF ground and excited states [42].
On the other hand, magnetic anisotropy in vdW honeycomb magnets with quenched orbital moments, such as and [42], is known to originate from anisotropic exchange interactions mediated by the SOC of ligand anions [42, 44]. Anisotropic exchange interactions through extended pathway (M-O-P-O-M) have also been demonstrated in honeycomb systems bridged by units [45]. In CCPS, the sublattice is linked by units, forming a comparable exchange pathway (Cr-S-P-S-Cr). This suggests that ligand-mediated anisotropic exchange may play a key role in determining the magnetic anisotropy. We therefore encourage further theoretical investigations, particularly DFT calculations including full SOC effects, to clarify the microscopic origin of magnetic anisotropy in this system.
The interlayer spacing in 2D vdW magnets has been identified as a crucial tuning parameter that controls the interlayer exchange coupling mediated by an M-X-X-M exchange pathway involving chalcogen -orbital overlap across the vdW gap [46]. A reduced interlayer spacing favors AFM interlayer, as Pauli repulsion in the overlap region dominates over the kinetic energy gain associated with interlayer hopping [46]. A similar mechanism, involving an interlayer Cr-S-S-Cr pathway, has also been proposed to explain the tuning of in CCPS [19]. In this context, the competition between these energy contributions provided a natural explanation of the observed reduction in the interlayer spacing below .
This temperature dependence, together with the field tuning of the interlayer spacing, emphasizes the presence of magnetoelastic coupling in CCPS. Such coupling is widely exploited in straintronics, where strain engineering enables effective control of a material’s magnetic and electronic properties [47, 48]. VdW materials are particularly attractive in this context due to their exceptional mechanical properties, including anisotropic compressibility [49]. For example, out-of-plane strain has been shown to significantly enhance interlayer AFM coupling in the vdW magnet [50, 51]. A linear relationship between and interlayer spacing under applied strain has also been reported in [52]. Therefore, the magnetoelastic coupling observed in this study points to a promising route for tuning the magnetic properties in CCPS. Recent studies on the isostructural have demonstrated control of polarization via out-of-plane strain, where local strain gradients drive ferroelectric domain formation through the flexoelectric effect [53, 54]. Such strain gradients are also expected to modulate the interlayer exchange coupling by altering the layer spacing and Cr-S covalent bonding. Accordingly, investigating strain-based control mechanism in CCPS presents both significant challenges and exciting opportunities, owing to the coupled tunability of its polarization and magnetic properties.
Conclusion
In summary, we have investigated the magnetic ground state of CCPS and its evolution under an applied magnetic fields using neutron scattering and magnetic measurements. The ordered moments in the “A-type” AFM structure are found to align along the axis. A spin-flop transition is observed when the field is applied along this direction, indicating weak in-plane uniaxial anisotropy. These results resolve the long-standing ambiguity regarding the magnetization orientation and pave the way for future studies of MEC mechanism in this system. A magnetic field applied along stabilizes a symmetry-adapted canted AFM state, accompanied by a pronounced magnetoelastic coupling effect. This coupling links the spin canting angle to the interlayer spacing, highlighting the intricate interplay between lattice and spin degrees of freedom. Our findings provide general guidelines for exploring similar phenomena in related 2D vdW multiferroics. Furthermore, out-of-plane strain is expected to modify the magnetic ordering temperature and spin configuration by tuning the interlayer spacing. Local strain gradient may enable simultaneous control of polarization and magnetism through the flexoelectric effect, opening promising opportunities for engineering multifunctional responses in low-dimensional materials.
Acknowledgement
This research used resources at the Spallation Neutron Source, a DOE Office of Science User Facility operated by the Oak Ridge National Laboratory. The beam time was allocated to CORELLI on proposal number IPTS-33430.1. Sample synthesis and characterization was performed at the Materials and Manufacturing Directorate at the Air Force Research Laboratory and were supported by the United States Air Force Office of Scientific Research (AFOSR), Grant LRIRs 23RXCOR003 and 26RXCOR010. J.-Q. Y. was supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Materials Sciences and Engineering Division.
This manuscript has been authored by UT-Battelle, LLC, under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for the United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/ downloads/doe-public-access-plan).
Supplementary Material: Magnetism and magnetoelastic effect in 2D van der Waals multiferroic CuCrPS
Jiasen Guo; Ryan P. Siebenaller; Michael A. Susner; Jiaqiang Yan; Zachary Morgan; Feng Ye
Crystal direction determination
Figure S1(a) shows the room temperature X-ray Laue pattern of the single crystal CCPS flake used in the magnetization measurements. The two-fold symmetry of the reciprocal space about the plane is clearly demonstrated for the monoclinic structure (space group ) with the -unique axis. In contrast, Figure S1(b) shows a typical X-ray Laue pattern lacking the two-fold symmetry, indicating coexisting structural domains. Such samples are coaligned for the neutron diffraction measurements.

Figure S1. Room temperature X-ray Laue pattern of CCPS with (a) single domain (b) multiple domains. The reciprocal lattice directions are labeled with denoting the projected component of onto the plane.
Simulation of the spin canting angle
The magnetic structure of CCPS can be approximated with a simplified model consisting AFM-coupled rigid FM layers stacking along the layer normal direction. The atomistic Hamiltonian of such a model is given by
where and are the spins representing the FM layers and and are the interlayer AFM exchange interaction and the coordination number. denotes the anisotropy constant for the uniaxial anisotropy along the direction. , , are the electron Laudé factor, Bohr magneton and magnetic field. Within the mean field description, the effective anisotropy field and exchange field are defined as
where for . Define the spin magnetic moment and the saturation magnetic moment , the Hamiltonian can be reformulated as
where and are the angles between the positive direction and the spin magnetic moments along unit vectors and . Figure S2 presents the simulated magnetization by minimizing Equation (3) using the parameters tabulated in Table 1 in the main text, in good agreement with the experimental data shown in Figure 1(c). The simulation reveals that for , the expression holds true when the is minimized. With this constraint enforced, an analytical expression, describes the field evolution of the spin canting angle . The simulated at K and 10 K as functions of magnetic field along the direction are plotted in Figure 4(b) in the main text.

Figure S2. Simulated magnetization (a) and (b) , respectively.
Structural domain volume fraction determination
The monoclinic structure of CCPS can be transformed into a pseudo-hexagonal structure via
where the subscript and denote for pseudo-hexagonal and monoclinic, and the transformation matrix
Accordingly, a given reflection can be transformed to via
A rotation operation converts the reflection via
where R is the rotation matrix in the pseudo-hexagonal structure. Putting Equation (6) and (S7) together and transforming the reflection back to the monoclinic frame, a reflection arising from a rotated structural domain can be indexed in the main domain frame via
with U being the effective transformation matrix given by
A reflection of the rotated domain could locate at fractional positions or integer positions, of which, the latter overlaps with a reflection of the main domain. Therefore, the domain volume fractions can be solved by reproducing the intensity of integer peaks. Neutron scattering intensities of a group of integer nuclear peaks collected at K was used for refining the domain volume fractions with the atomic positions and the displacement parameters refined from single crystal x-ray data collected at the same temperature [Table S1]. Table S2 presents U of the rotated domains and the corresponding volume fractions refined using JANA2020 [1]. The goodness-of-fit is shown in Figure S3.
| Main domain | 60° domain | 120° domain | 180° domain | 240° domain | 300° domain | |
| U | ||||||
| Volume fraction | 0.424 | 0.003 | 0.054 | 0.461 | 0.020 | 0.038 |
Table S2. The transformation matrices U from the main domain to rotated domains and the corresponding refined volume fractions.

Figure S3. vs at K for the integer nuclear peaks used for the determination of the domain volume fractions.
| S | 0.75165(8) | 0.18486(5) | 0.62578(3) | 0.0049(2) | 0.0077(2) | 0.0082(2) | -0.0025(1) | 0.0032(2) | -0.0042(1) |
| S | 0.73167(8) | 0.17470(4) | 0.12569(3) | 0.0061(2) | 0.0065(2) | 0.0072(2) | -0.0026(1) | 0.0029(2) | -0.0031(1) |
| S | 0.78025(8) | 0.49532(4) | 0.62364(4) | 0.0050(2) | 0.0037(2) | 0.0087(2) | -0.0008(1) | -0.0027(2) | 0.0006(1) |
| P | -0.05303(8) | 0.33185(4) | 0.16570(4) | 0.0040(2) | 0.0044(2) | 0.0041(2) | 0.0003(1) | 0.0002(2) | 0.0002(1) |
| Cr | 0 | 0.66466(3) | 0.25 | 0.0028(2) | 0.0034(2) | 0.0037(2) | 0 | -0.0001(1) | 0 |
| Cu | -0.0652(2) | 0.00221(7) | 0.6454(1) | 0.0101(2) | 0.0064(2) | 0.0611(4) | 0.0000(1) | 0.0169(2) | 0.0002(2) |
Table S1. The atomic positions and displacement parameters in space group for CCPS at K refined from single crystal x-ray data and used in the determination of domain volume fractions. All quantities are in fractional coordinates.
Magnetic structure refinement
The multi-domain magnetic structure refinement was carried out with an in-house developed python script. A group of integer nuclear peaks at K was identified to determine the scale factor (a factor of 0.875 was applied for the peaks to correct for the misalignment of g out of the total 8 g sample). For each peak, the criteria of determining the contributing domains is that they would assure integer peaks upon transforming the original peak to specific domains. Figure S4 demonstrate the vs for the scale factor determination where the atomic coordinates and displacement parameters are given in Table S3, refined from single crystal x-ray data collected at K (lowest accessible ). Magnetic intensities were obtained by subtracting the K data from the K data for a list of integer peaks and the contributing domains determined following the same criteria.

Figure S4. vs for the (a) nuclear and (b) magnetic peaks. (c) The residual as the moments cant away from the direction. (d) Illustration of the canting angle definition.
| Cr | 0.1304(2) | -0.1651(2) | -0.6438(1) | 0.0066(5) | 0.0077(7) | 0.0096(5) | -0.0030(5) | 0.0006(4) | -0.0009(5) |
| Cr | -0.3724(2) | -0.6645(2) | -0.6427(1) | 0.0082(5) | 0.0078(7) | 0.0076(5) | -0.0030(5) | 0.0014(4) | -0.0009(5) |
| Cu | -0.3033(2) | -0.9982(1) | -0.5363(1) | 0.0124(5) | 0.0108(6) | 0.0422(8) | -0.0009(5) | 0.0127(5) | 0.0002(6) |
| Cu | 0.0611(2) | -0.4982(1) | -0.7509(1) | 0.0128(5) | 0.0102(6) | 0.0378(7) | -0.0007(5) | 0.0080(5) | 0.0011(6) |
| S | -0.1284(4) | -0.3078(3) | -0.7720(2) | 0.0079(9) | 0.008(1) | 0.0083(9) | 0.0009(8) | 0.0016(7) | 0.0014(9) |
| S | -0.1465(4) | -0.6806(3) | -0.7693(2) | 0.0060(8) | 0.008(1) | 0.0074(8) | -0.0018(8) | 0.0000(6) | -0.0001(8) |
| S | -0.6383(5) | -0.1693(3) | -0.7670(2) | 0.0087(9) | 0.008(1) | 0.0104(9) | 0.0004(8) | 0.0035(7) | 0.0021(8) |
| S | -0.6157(4) | -0.8214(3) | -0.7638(2) | 0.0082(9) | 0.010(1) | 0.0096(9) | 0.0001(8) | 0.0016(7) | -0.0019(8) |
| S | -0.0945(4) | -0.0029(3) | -0.7642(2) | 0.0098(7) | 0.010(1) | 0.0147(9) | -0.0008(8) | 0.0001(7) | 0.0000(8) |
| S | -0.5802(4) | -0.5076(3) | -0.7759(2) | 0.0070(8) | 0.012(1) | 0.012(1) | 0.0021(8) | 0.0025(7) | 0.0006(9) |
| P | -0.4324(4) | -0.3291(3) | -0.7300(2) | 0.0119(9) | 0.008(1) | 0.008(1) | 0.0016(9) | 0.0033(8) | 0.0002(8) |
| P | 0.0823(4) | -0.8352(3) | -0.7265(2) | 0.009(1) | 0.013(1) | 0.0046(9) | -0.0008(9) | 0.0000(8) | 0.0013(8) |
| S | -0.1403(4) | -0.5032(3) | -0.5233(2) | 0.0067(7) | 0.0057(9) | 0.0078(9) | 0.0010(7) | -0.0013(6) | -0.0009(8) |
| S | 0.3914(5) | -0.6698(2) | -0.5203(2) | 0.0102(9) | 0.009(1) | 0.0102(9) | -0.0030(8) | 0.0019(8) | -0.0009(8) |
| S | 0.3451(4) | -0.0063(3) | -0.5127(2) | 0.0075(7) | 0.0054(9) | 0.0095(8) | 0.0016(7) | -0.0005(6) | 0.0011(7) |
| S | 0.3688(4) | -0.3221(3) | -0.5233(2) | 0.0084(9) | 0.012(1) | 0.012(1) | 0.0029(8) | 0.0043(8) | 0.0026(9) |
| S | -0.1181(4) | -0.8092(3) | -0.5151(2) | 0.0086(9) | 0.009(1) | 0.0121(9) | 0.0028(8) | 0.0032(7) | 0.0037(9) |
| S | -0.0975(5) | -0.1809(3) | -0.5176(2) | 0.0099(9) | 0.010(1) | 0.0124(9) | -0.0024(8) | 0.0036(7) | -0.0026(8) |
| P | -0.3253(4) | -0.3357(2) | -0.5618(2) | 0.0062(8) | 0.004(1) | 0.012(1) | -0.0004(8) | 0.0018(8) | 0.0010(8) |
| P | 0.1876(4) | -0.8283(3) | -0.5581(2) | 0.0034(9) | 0.008(1) | 0.0096(9) | 0.0015(8) | -0.0008(8) | 0.0002(8) |
Table S3. The atomic positions and displacement parameters in space group for CCPS at K refined from single crystal x-ray data and used in the magnetic structure refinement. All quantities are in fractional coordinates.
We first performed representational analysis was performed for the parent structure [space group ] with a propagation vector using SARAh [2]. The magnetic representation can be decomposed into , listed in Table S4. corresponds to magnetic space group , which allows for FM along the direction and A-type antiferromagnetism (AFM) along the and directions. On the contrary, corresponds to magnetic space group , allowing FM along and directions and A-type AFM along the direction. A constrained refinement with collinear moments with equal size on the two independent Cr sites was then performed. The refined magnetic structure (AFM) is described by magnetic space group with A-type AFM along the direction and the ordered moment size (Figure S4). Next, within this magnetic space group, FM components along the and directions are introduced, which increases the residual,
| component | component | ||||||||||
| IR | BV | Atom | IR | BV | Atom | ||||||
| () | Cr1 | 1 | 0 | 0 | () | Cr1 | 1 | 0 | 0 | ||
| Cr2 | -1 | 0 | 0 | Cr2 | 1 | 0 | 0 | ||||
| Cr1 | 0 | 1 | 0 | Cr1 | 0 | 1 | 0 | ||||
| Cr2 | 0 | 1 | 0 | Cr2 | 0 | -1 | 0 | ||||
| Cr1 | 0 | 0 | 1 | Cr1 | 0 | 0 | 1 | ||||
| Cr2 | 0 | 0 | -1 | Cr2 | 0 | 0 | 1 | ||||
| Cr3 | 1 | 0 | 0 | Cr3 | 1 | 0 | 0 | ||||
| Cr4 | -1 | 0 | 0 | Cr4 | 1 | 0 | 0 | ||||
| Cr3 | 0 | 1 | 0 | Cr3 | 0 | 1 | 0 | ||||
| Cr4 | 0 | 1 | 0 | Cr4 | 0 | -1 | 0 | ||||
| Cr3 | 0 | 0 | 1 | Cr3 | 0 | 0 | 1 | ||||
| Cr4 | 0 | 0 | -1 | Cr4 | 0 | 0 | 1 |
Table S4. Basis vectors (BVs) for the space group with magnetic propagation vector . The decomposition of the magnetic representation is . The two magnetic Cr atoms sites locate at Wyckoff position with fractional coordinates: Cr1:(0.14724, 0.33531, 0.39605), Cr3:(0.64443, 0.83510, 0.39721) with Cr2 and Cr4 generated by the c-gliding operation (X, -Y+1, Z+1/2), respectively.
as shown in Figure S4. In the above equation, and are calculated and observed peak intensities. Thus, within the resolution of the neutron diffraction data, we are convinced that the ground state ordered moments in CCPS align along the direction. Table S5 tabulates the observed and calculated intensities. For comparison, the calculated intensity of reported models in the literature are also included with .
| Reflection | Observation | AFM | AFM | AFM |
| 31 (3) | 27.30 | 27.30 | 27.30 | |
| 28 (5) | 27.30 | 27.30 | 27.30 | |
| 22 (1) | 21.20 | 21.20 | 21.20 | |
| 22 (2) | 21.20 | 21.20 | 21.20 | |
| 5 (0) | 5.40 | 2.09 | 2.13 | |
| 6 (0) | 5.95 | 2.31 | 7.32 | |
| 6 (0) | 5.95 | 2.31 | 7.32 | |
| 6 (1) | 5.40 | 2.09 | 2.13 | |
| 2 (0) | 1.89 | 1.09 | 1.10 | |
| 2 (0) | 2.08 | 1.21 | 2.42 | |
| 5 (1) | 5.46 | 13.91 | 7.57 | |
| 5 (1) | 5.46 | 13.91 | 7.57 | |
| 5 (1) | 5.46 | 13.91 | 7.57 | |
| 8 (1) | 5.46 | 13.91 | 7.57 | |
| 5 (1) | 4.97 | 3.65 | 3.66 | |
| 5 (1) | 5.49 | 4.03 | 6.05 | |
| 5 (1) | 6.09 | 3.18 | 7.19 | |
| 5 (1) | 5.54 | 2.89 | 2.92 | |
| 2 (0) | 2.07 | 1.44 | 2.31 | |
| 2 (1) | 1.89 | 1.31 | 1.32 | |
| 0 (0) | 0.00 | 15.77 | 3.94 | |
| 0.07 | 0.34 | 0.17 |
Table S5. Comparison of observed and calculated magnetic intensities for three A-type AFM models with moments along (AFM, this work), (AFM [3, 4]) and plane diagonal (AFM [5]) directions. Calculation of the residual excludes the non-observed peak.
Extracting spin canting angle from fitting magnetic peak intensity
In the simplified model considering rigid FM layers AFM-coupled along the layer normal, an empirical model [6]
can be used to correlate the canting angle between the FM moment and the direction with the magnetic field strength along the direction, where , , are constants. Magnetic peak intensity is subsequently determined from the canting angle. Figure S5 presents the field evolution of the magnetic peak intensity at various temperatures [ K in Figure 3(a)], simultaneously fitted with the empirical model, which allows us to extracted the canting angle as a function of field, plotted in Figure 3(b).

Figure S5. Field evolution of the magnetic peak (0, 0, 1) intensity at various temperatures. Solid lines are empirical model fits based on Eq. (11).
References
- [1]V. Petříček, L. Palatinus, J. Plášil, and M. Dušek, Jana2020 – a new version of the crystallographic computing system Jana, Z. Kristallogr. Cryst. Mater. 238, 271 (2023).DOI
- [2]A. S. Wills, A new protocol for the determination of magnetic structures using simulated annealing and representational analysis (SARAh), Physica B: Condens. Matter 276-278, 680 (2000).DOI
- [3]C. B. Park, A. K. T. Kim, D. R. Patil, S. A. Guda, N. Ter-Oganessian, and K. H. Kim, Observation of spin-induced ferroelectricity in a layered van der Waals antiferromagnet CuCrP₂S₆, Adv. Electron. Mater. 8, 2101072 (2022).DOI
- [4]X. Wang, Z. Shang, C. Zhang, J. Kang, T. Liu, X. Wang, S. Chen, H. Liu, W. Tang, Y. Zeng, J. Guo, Z. Cheng, L. Liu, D. Pan, S. Tong, B. Wu, Y. Xie, G. Wang, J. Deng, T. Zhai, H. Deng, J. Hong, and J. Zhao, Electrical and magnetic anisotropies in van der waals multiferroic CuCrP₂S₆, Nat. Commun. 14, 840 (2023).DOI
- [5]V. B. Cajipea, J. Ravez, V. Maisonneuve, A. Simon, C. Payen, R. Von der Mühll, and J. Fischer, Copper ordering in lamellar CuMP₂S₆ (M = Cr, In): Transition to an antiferroelectric or ferroelectric phase, Ferroelectrics 185, 135 (1996).DOI
- [6]F. Ye, M. Matsuda, Z. Morgan, T. Sherline, Y. Ni, H. Zhao, and G. Cao, Magnetic structure and spin fluctuations in the colossal magnetoresistance ferrimagnet Mn₃Si₂Te₆, Phys. Rev. B 106, L180402 (2022).DOI