betaFeedback

WorksPhysics and Astronomy

Non-coplanar helimagnetism in the layered van-der-Waals metal DyTe3_3

Abstract

Magnetic materials with highly anisotropic chemical bonding can be exfoliated to realize ultrathin sheets or interfaces with highly controllable optical or spintronics responses, while also promising novel cross-correlation phenomena between electric polarization and the magnetic texture. The vast majority of these van-der-Waals magnets are collinear ferro-, ferri-, or antiferromagnets, with a particular scarcity of lattice-incommensurate helimagnets of defined left- or right-handed rotation sense, or helicity. Here we use polarized neutron scattering to reveal cycloidal, or conical, magnetic structures in DyTe3_3, with coupled commensurate and incommensurate order parameters, where covalently bonded double-slabs of dysprosium square nets are separated by highly metallic tellurium layers. Based on this ground state and its evolution in a magnetic field as probed by small-angle neutron scattering (SANS), we establish a one-dimensional spin model with off-diagonal on-site terms, spatially modulated by the unconventional charge order in DyTe3_3. The CDW-driven term couples to antiferromagnetism, or to the net magnetization in applied magnetic field, and creates a complex magnetic phase diagram indicative of competing interactions in an easily cleavable helimagnet. Our work paves the way for twistronics research, where helimagnetic layers can be combined to form complex spin textures on-demand, using the vast family of rare earth chalcogenides and beyond.

  • These two authors contributed equally to this work

† [email protected]

‡ Present address: Department of Physics, Faculty of Mathematics and Natural Sciences, Universitas Indonesia, Depok 16424, Indonesia.

§ [email protected]

Magnetism in layered materials, held together by weak van-der-Waals interactions, is an active field of research spurred on by the discovery of magnetic ordering in monolayer sheets of ferromagnets and antiferromagnets [1, 2, 3, 4]. At the frontier of this field, helimagnetic layered systems, where magnetic order has a fixed, left- or right-handed rotation sense, have been predicted to host complex spin textures [5, 6] and to serve as controllable multiferroics platforms, where magnetic order is readily tuned by electric fields or currents [7, 8, 9, 10]. However, most layered van-der-Waals magnets are commensurate ferro-, antiferro-, or ferri-magnets [3, 4]; the rare helimagnets provided to us by nature are often modulated along the stacking direction, with relatively simple spin arrangement in individual layers (Table E1).

CompoundSpace groupq\mathbf{q}-vectorTransportMagnetismRef.
DyTe3_3CmcmCmcmqcyc∼(0,b∗,0.207c∗)\mathbf{q}_{\mathrm{cyc}} \sim(0,b^*,0.207c^*)MNCPthis work
DyTe3_3qAFM∼(0,b∗,0.5c∗)\mathbf{q}_{\mathrm{AFM}} \sim(0,b^*,0.5c^*)
Fe5−x_{5-x}GeTe2_2Rˉ3m\bar{R}3m±13(1,1,3)\pm\frac{1}{3}(1,1,3)MNCP[51, 52, 53]
Fe5−x_{5-x}GeTe2_2±310(0,0,3)\pm\frac{3}{10}(0,0,3)
AgCrSe2_2(0.037,0.037,3/2)(0.037,0.037,3/2)ICP[54, 55]
NiI2_2(0.138,0,1.457)(0.138,0,1.457)MITCP[56, 57]
NiBr2_2(0.027,0.027,3/2)(0.027,0.027,3/2)ICP[58, 59, 60]
CoI2_2Pˉ3m1\bar{P}3m1(1/12,1/12,1/2)(1/12,1/12,1/2)-CP[56]
CoI2_2(1/8,0,1/2)(1/8,0,1/2)
MnI2_2(0.181,0,0.439)(0.181,0,0.439)ICP[60, 61]
Co1/3_{1/3}NbS2_2P6322P6_3 22(0.5,0,0)(0.5,0,0)MNCP[62, 63, 64]
Co1/3_{1/3}TaS2_2(0.5,0,0)(0.5,0,0)MNCP[65]
Cr1/3_{1/3}NbS2_2(0,0,0.025)(0,0,0.025)MNCP[66, 67, 68]
Cr1/3_{1/3}TaS2_2(0,0,0.081)(0,0,0.081)MNCP[66, 69, 70, 71]

Table E1. Magnetic properties of van-der Waals systems with complex magnetic order. Electrical transport properties are categorized into metals (M), insulators (I), and materials with metal-to-insulator transition (MIT), and spin textures are classified into coplanar (CP) and noncoplanar (NCP). DyTe3_3 is the only metallic compound that is (a) incommensurate with the underlying lattice, and (b) has a component of the modulation vector q\mathbf{q} perpendicular to the stacking direction. Moreover, it is a rare example of noncoplanar (NCP) magnetism in a bulk vdW compound.

In the quest for helimagnetism in layered structures with weak van-der-Waals bonds, we focus on rare earth tritellurides RTe3\mathit{R}\mathrm{Te}_{3} (R\mathit{R}: rare earth element). These materials form a highly active arena of research regarding the interplay of correlations and topological electronic states [11, 12, 13, 14, 15]. Their structure, which can be exfoliated down to the thickness of a few monolayers [16, 17], is composed of tellurium Te2\mathrm{Te}_{2} double-layers and covalently bonded RTe\mathit{R}\mathrm{Te} slabs, with characteristic square net motifs in both (Figure 1a) [18]. Tellurium 5p5p electrons are highly localized in Te2\mathrm{Te}_{2} square net bilayers (acac plane), in which they form highly dispersive bands with elevated Fermi velocity [16, 19, 20, 21]. This quasi two-dimensional electronic structure amplifies correlation phenomena, such as the formation of charge-density wave (CDW) order [22] and superconductivity [23, 24], while also hosting protected band degeneracies [16, 25]. In view of intense research efforts on RTe3\mathit{R}\mathrm{Te}_{3}, it is remarkable that their magnetism has never been discussed in detail; in particular, no full refinement of magnetic structures is available [26, 27, 28, 29, 30, 31].

Image not extracted

Figure 1. Conical helimagnetism in the layered square-lattice antiferromagnet DyTe3DyTe_3. a, Crystallographic unit cell with covalently bonded DyTe bilayers and metallic Te bilayers, with natural cleaving plane (dashed). b, Magnetic exchange interactions in a single DyTe double-square net bilayer. c, Zigzag chain illustration of double-square net structure in DyTe3DyTe_3. The interactions J1J_1, J2J_2 connect nearest and next-nearest neighbours in the zigzag chain model, respectively; but the inter-atomic distance in the full crystallographic structure is shorter for J2J_2, and its antiferromagnetic coupling strength is dominant. Conical, noncoplanar helimagnetism is resolved in the zero-field ground state by neutron scattering. The cone direction, parallel to the aa-axis, alternates both between pairs of magnetic sites in a zigzag chain, and between stacked zigzag chains. This texture causes polarization along the bb-axis, i.e. perpendicular to square net bilayers (yellow arrow). Note: The full magnetic unit cell extends two times further along the chain direction (Figure E1). d, Weak anisotropy of the magnetic susceptibility χ\chi in DyTe3DyTe_3. The softest direction is H∥cH \parallel c, consistent with the modulation direction of the magnetic order in panel c. The inset shows Curie-Weiss temperatures and temperatures TχT_\chi of maximal χ\chi, measured in a small magnetic fields along three crystallographic directions.

Here, we report on helimagnetic, cone-type orders of DyTe3\mathrm{DyTe}_{3} using polarized elastic neutron scattering. We reveal the magnetic texture in real space, probe its evolution with temperature and magnetic field, and reveal its relationship to charge-density wave formation. In DyTe3\mathrm{DyTe}_{3}, dysprosium moments are arranged in square net bilayers, where each ion has neighbours within its own layer, and within the respective other layer (Figure 1b). As all zero-field magnetic orders of DyTe3\mathrm{DyTe}_{3} are uniform along the crystallographic aa-axis, it is reasonable to understand each square net bilayer as an effective zigzag chain of magnetic rare earth ions and to define magnetic interactions J1J_{1} and J2J_{2} in terms of nearest- and next-#+#+#+#+ nearest neighbours on the zigzag chain, respectively. On such chains, our experiment shows that pairs of ions have cones pointing along the same direction, followed by a flip of the cone axis (Figure 1c, which illustrates half a magnetic unit cell). The coupling between two DyTe bilayers, i.e. between two zigzag chains, is antiferromagnetic. Despite this complex cone arrangement, the magnetic structure defines a fixed sense of rotation, or helicity. Our theoretical spin model shows that charge-density wave order in rare earth tellurides causes local symmetry breaking, allows for off-diagonal on-site coupling terms in the Hamiltonian, and drives lattice-incommensurate magnetism when combined with antiferromagnetic interactions or with a net magnetization. We further discuss magnetocrystalline anisotropy in this layered structure, with an unconventional combination of metallic and covalent bonds. Helimagnetism of Dy rare earth moments with 4f94f^9 magnetic shell emerges despite the naive expectation of strong preference for easy-axis or easy-plane anisotropy for 2S+1LJ=6H15/2{}^{2S+1}L_J = {}^6H_{15/2}, with large orbital angular momentum L=5L = 5.

Magnetic properties of DyTe3\mathrm{DyTe}_3

Some essential magnetic properties of DyTe3\mathrm{DyTe}_3 are apparent already from the magnetic susceptibility χ\chi in Figure 1d. In the high temperature regime, anisotropy in the Curie-Weiss law indicates easy-plane behaviour of magnetic moments, favouring the acac plane with uniaxial anisotropy constant K1=22(1) kJ m−3K_1 = 22(1)\ \mathrm{kJ\,m^{-3}} (Methods). At low temperatures, the strongest enhancement of χ\chi occurs when the magnetic field H\mathbf{H} is along the cc-axis, i.e. parallel to the zigzag direction defined in Figure 1c. We may deduce that the magnetic moments are aligned, predominantly, along the aa and bb axes. All susceptibility curves show maxima around Tχ=4.5 KT_\chi= 4.5\,\mathrm{K}, quite far above the onset of three-dimensional, long-range magnetic order, as shown in the following.

We characterize the phase transition in DyTe3\mathrm{DyTe}_3 using thermodynamic and transport probes in Figure 2a,Figure 2b. The specific heat C(T)C(T) shows a two-peak anomaly, describing the transitions from the paramagnetic (PM) regime to phase II at TN2=3.85 KT_{\mathrm{N2}} = 3.85\,\mathrm{K} and to phase I at TN1=3.6 KT_{\mathrm{N1}} = 3.6\,\mathrm{K}. Below TN1T_{\mathrm{N1}}, the resistivities in the acac basal plane drop abruptly, suggesting a clear correlation between the behaviour of freely moving conduction electrons and the magnetic structure. The presence of a partial charge gap in the electronic structure, related to magnetic ordering, is inferred from an increase of the ratio of resistivities ρa\rho_a and ρc\rho_c. Simultaneously, as discussed in the following, strong neutron scattering intensity appears below TN2T_{\mathrm{N2}} at two independent positions in reciprocal space, c.f. Figure 2c. The magnetic scattering intensity rises abruptly upon cooling below TN2T_{\mathrm{N2}}.

Image not extracted

Figure 2. Bulk characterization and magnetic structure of DyTe3_3 from elastic neutron scattering. a, Specific heat showing two transitions at TN1T_{\mathrm{N1}} and TN2T_{\mathrm{N2}}, with reference data from the nonmagnetic analogue LaTe3_3 [32]. b, Below TN1T_{\mathrm{N1}}, where resistivity in the acac basal plane drops sharply, significant anisotropy ρc/ρa\rho_c/\rho_a indicates (partial) gapping of electronic states along the cc-axis. c, Temperature dependence of magnetic scattering intensity for antiferromagnetic qAFM\mathbf{q}_{\mathrm{AFM}} and cycloidal qcyc\mathbf{q}_{\mathrm{cyc}} reflections, with Miller indices (HKL)=(0,1,0.5)(HKL) = (0, 1, 0.5) and (0,1,0.207)(0, 1, 0.207), respectively. Grey and green background shading mark phases I and II, respectively. d, Experimental geometry for polarized neutron scattering from layered DyTe3_3 (grey plane, square net of Dy indicated). The scattering plane (blue) is spanned by wavevectors ki\mathbf{k}_i, kf\mathbf{k}_f of incoming and outgoing neutron beams, respectively. Separating spin-flip (SF, red) and non-spin flip (NSF, blue) scattering intensities at the detector, we identify conical magnetic order (top). e, Linescan in momentum space, with highlights for three types of reflections including a weak higher harmonic corresponding to 3×qcyc3 \times\mathbf{q}_{\mathrm{cyc}}. f, Simulation of magnetic intensity from magnetic structure model, corresponding to the linescan in panel e (Methods).

To obtain this neutron data, a single-domain crystal of DyTe3_3 is mounted on an aluminium holder and is pre-aligned by means of Laue x-ray diffraction. More quantitatively, we determine the crystallographic directions in DyTe3_3 using the crystallographic extinction rule (Extended Section, Figure E11). Figure 2d describes the geometry of our neutron scattering experiment. The scattering plane that includes the incoming and outgoing neutron beams ki\mathbf{k}_{i} and kf\mathbf{k}_{f}, is spanned by the bb- and cc-axes. Hence, reflections of the type Q=kf−ki\mathbf{Q} = \mathbf{k}_{f} - \mathbf{k}_{i} with Miller indices (0KL)(0KL) can be detected, as in Figure 2e, where a line scan along (01L)(01L) provides sharp magnetic intensity. Three types of magnetic peaks Q=G+q\mathbf{Q} = \mathbf{G} + \mathbf{q}, with G\mathbf{G} a reciprocal lattice vector, are observed: A commensurate (C) reflection qAFM=(0,b∗,qAFM)\mathbf{q}_{\mathrm{AFM}} = (0,b^{*},q_{\mathrm{AFM}}), qAFM=0.5c∗q_{\mathrm{AFM}} = 0.5c^{*}; an incommensurate (IC) reflection qcyc=(0,b∗,qcyc)\mathbf{q}_{\mathrm{cyc}} = (0,b^{*},q_{\mathrm{cyc}}), qcyc=0.207c∗q_{\mathrm{cyc}} = 0.207c^{*}, where b∗=2π/bb^{*} = 2\pi/b and c∗=2π/cc^{*} = 2\pi/c are reciprocal lattice constants (Methods). There is also a higher harmonic (3Q) reflection, corresponding to three times the length of qcyc\mathbf{q}_{\mathrm{cyc}}, which describes an anharmonic distortion of the texture. As a main result of this work, we ascribe qcyc\mathbf{q}_{\mathrm{cyc}} to a cycloidal structure in the magnetic ground state of DyTe3_3, that results from a coupling qAFM±qCDW\mathbf{q}_{\mathrm{AFM}} \pm\mathbf{q}_{\mathrm{CDW}} between the C order and a charge-density wave (CDW) modulation qCDW\mathbf{q}_{\mathrm{CDW}} at (0,0,qCDW)(0,0,q_{\mathrm{CDW}}), qCDW=0.29c∗q_{\mathrm{CDW}} = 0.29c^{*} in the rare earth tritelluride family.

Four-panel confirmation of crystallographic extinction rule: panels a and c show strong nuclear Bragg reflections at (020) and (002), respectively, while panel b shows zero intensity at (031); panel d shows the reciprocal-space diagram with strong nuclear Bragg reflections, the first Brillouin zone, and reflections (000), (020), (040), and (002)

Figure E11. Confirmation of crystallographic extinction rule in DyTe3\mathrm{DyTe}_{3} in sample A. Reflections h+k=oddh+k=\text{odd} are forbidden in Cmcm\mathit{Cmcm} (space group 63), or specifically k=2nk=2n is required in the (0KL)(0KL) scattering plane, which is used for the present experiment. We observe zero intensity at (031) and 800 counts per second (cps) at (021); the latter reflection is not shown in this figure. When instead rotating the sample into the (HK0)(HK0) scattering plane, (130) has 1200 cps and (120) exhibits zero intensity. Specifically, the comparison of (130) and (031) allows us to confirm the alignment of the crystal in our neutron experiment.

Ground state magnetic structure model

We reveal the helimagnetic structure in the ground state of DyTe3_3 using polarized neutron scattering. As shown in Figure 2d, the incident neutron spins were polarized perpendicular to the scattering plane. We employ a magnetized single-crystal analyzer to select the energy and spin state of the scattered neutrons (Methods). The scattering processes in which the neutron spins are reversed (remain unchanged) is referred to as spin-flip, SF (non-spin-flip, NSF). For SF scattering, it is required that magnetic moments m\mathbf{m} have a component perpendicular to the spin of the incoming neutron. This means that SF and NSF scattering detect components of m\mathbf{m} within (mb,mc)(m_b,m_c) and perpendicular to (ma)(m_a) the scattering plane, respectively.

Polarization analysis of the magnetic reflections shows that qcyc\mathbf{q}_{\mathrm{cyc}} and qAFM\mathbf{q}_{\mathrm{AFM}} relate to different vector components of the ordered magnetic moment (Figure 3a,Figure 3b and Figure 3e,Figure 3f). We find no hint of SF scattering at qAFM\mathbf{q}_{\mathrm{AFM}}, demonstrating collinear antiferromagnetism with magnetic moments exclusively along the aa-direction. The incommensurate part qcycq_{\mathrm{cyc}}, in contrast, has no NSF intensity and roughly equal SF signals at various positions in reciprocal space (Figure 3e,Figure 3f and insets). As neutron scattering detects the part of m\mathbf{m} that is orthogonal to Q\mathbf{Q}, comparison of magnetic reflections situated at nearly orthogonal directions in momentum space suggests mbm_b and mcm_c components are both finite in the ground state.

Two components of the cone-type order in the ground state of DyTe3

Figure 3. Two components of the cone-type order in the ground state of DyTe3_3. In the geometry of Figure 2d, non spin-flip (NSF) and spin-flip (SF) neutron scattering measure the magnetic moment along the crystallographic aa-axis and in the bcbc plane, respectively. a,b, Polarization analysis of antiferromagnetic component qAFM\mathbf{q}_{\mathrm{AFM}}: absence of spin flip (SF) intensity indicates absence of mbm_b and mcm_c, while non-spin flip (NSF) intensity reveals dominant mam_a. c,d, Two views of the antiferromagnetic collinear component derived from this data, confirmed by full refinement of a number of magnetic reflections in Figure E8. e,f, Polarization analysis for the incommensurate cycloidal reflection qcyc\mathbf{q}_{\mathrm{cyc}}. g, Derived magnetic structure model for qcyc\mathbf{q}_{\mathrm{cyc}}, where mam_a (NSF) vanishes while mcm_c and mbm_b (SF) both appear at comparable magnitudes in phase I. The index jj labels four dysprosium atoms in four layers within the chemical unit cell (red box), where the cycloids at j=2,4j = 2, 4 are phase-shifted with respect to cycloids at j=1,3j = 1, 3. The numerical value of δ\delta is determined in § 6.6. As for a,b,e, and f, we have subtracted background signals, and corrected the effect of imperfect beam polarization.

We determine the quantitative relationship between magnetic moments within a DyTe bilayer (within an effective zigzag chain), by comparing the observed and calculated magnetic structure factors under the constraints imposed by polarized neutron scattering, c.f. Figure E8. The analysis for qAFM\mathbf{q}_{\mathrm{AFM}} demonstrates up-up-down-down type ordering along the zigzag chain, visualized from two perspectives in Figure 3c,Figure 3d. At qcyc\mathbf{q}_{\mathrm{cyc}}, the refinement yields a cycloid with a phase delay δ\delta between the upper and lower sheets in a zigzag chain, see Figure 3g. In effect, pairs of nearly parallel magnetic moments are followed by a significant rotation of the moment direction. The coupling between zigzag chains is antiferromagnetic, as imposed by the Miller index K=1K = 1 (qb=b∗q_b = b^*) component in both qcyc\mathbf{q}_{\mathrm{cyc}} and qAFM\mathbf{q}_{\mathrm{AFM}}. Superimposing the three components mam_a, mbm_b, and mcm_c, we realize the noncoplanar, helimagnetic cone texture of Figure 1c that is, to our knowledge, unique in both insulators and metals. In Extended Sections E3 A, E4 A, we discuss the presence of magnetic domains in the sample and how the occurrence of higher harmonic reflections further supports our magnetic structure model.

Figure E8. Magnetic structure refinement for the ground state of DyTe3_3. Antiferromagnetic (commensurate, qAFM\mathbf{q}_{\mathrm{AFM}}) and cycloidal (incommensurate qcyc\mathbf{q}_{\mathrm{cyc}}) components are refined separately from neutron scattering data, without polarization analyzer, at T=2.2T = 2.2 Ksample B. a, Averaging two AFM domains with equal weight, we find good agreement [reliability factor R=0.0578R = 0.0578] between the measured neutron structure factor FobsF_{\mathrm{obs}} and model calculations FcalF_{\mathrm{cal}}. b, Likewise, good agreement is found when summing two δ\delta-domains for the incommensurate order qcyc\mathbf{q}_{\mathrm{cyc}}. Here, we minimized RR by adjusting the value of the phase shift δ\delta between sheets in a bilayer, as defined in Figure 3g, and the elliptic distortion Y/XY/X of the cycloid (§ 6.6).

Charge density wave and magnetic order

We now argue that cone-type magnetism in DyTe3DyTe_3 is realized through (i) a spatial modulation of near-neighbour exchange interactions J1J_1, J2J_2 in presence of charge-density wave (CDW) order and (ii) unconventional single-ion anisotropy. We turn first to (i), that is the role of the CDW in stabilizing noncoplanar helimagnetism in DyTe3DyTe_3. We use a 1D chain model to reproduce key features of the modulated magnetic order, neglecting the material’s three-dimensionality (Methods).

In DyTe3DyTe_3, the local environment and bond characteristics of dysprosium ions in a DyTe square net bilayer (in a zigzag chain) are spatially modulated by the CDW in the adjacent Te2Te_2 sheets, c.f. Figure 4b [11, 12, 13, 14, 18, 22, 23, 24, 31, 32, 33]. The simplest model approach is to decouple the zigzag chain, with two atoms per unit cell, into two one-dimensional chains, with one atom per unit cell. This allows for a two-parameter model, built from Ising-like exchange interactions together with a spatially modulated onsite coupling,

Modeling helimagnetism in DyTe3

Figure 4. Modeling helimagnetism in DyTe3. a, Magnetic susceptibility χ\chi for H∥[101]\mathbf{H} \parallel[101], overlaid with phase boundaries from specific heat (white open circles) and magnetization (black circles): I (conical ground state), II, III, IV, and V. b, Charge density wave (CDW) on tellurium layers [18], where orange (grey) spheres are distorted (undistorted) ionic positions. Coupling of Dy (violet) and CDW drives lattice-incommensurate magnetic order through spatially modulated modification of the local environment of Dy-ions. c, Squared moment amplitudes IAFM=∣Sa(q=0.5c∗)∣2I_{\mathrm{AFM}} = \lvert S^{a}(q = 0.5c^{*}) \rvert^{2} (blue) and Icyc=∣Sb(q=0.5c∗±qCDW)∣2+∣Sc(q=0.5c∗±qCDW)∣2I_{\mathrm{cyc}} = \lvert S^{b}(q = 0.5c^{*} \pm q_{\mathrm{CDW}}) \rvert^{2} + \lvert S^{c}(q = 0.5c^{*} \pm q_{\mathrm{CDW}}) \rvert^{2} (red) as functions of ECDW/J2AFME_{\mathrm{CDW}}/J_{2}^{\mathrm{AFM}}, where ECDW=ECDWab=ECDWacE_{\mathrm{CDW}} = E_{\mathrm{CDW}}^{ab} = E_{\mathrm{CDW}}^{ac}, from model calculations according to Eq. (1); five and ten percent threshold for the incommensurate part indicated by black and grey dashed lines (Methods). d, Local environment of Dy in DyTe3, with covalent (metallic) bonds to Te-B (Te-A) depicted by solid (dashed) lines, respectively. e, Charge density (CD) of Dy 4f94f^{9} shell under the influence of crystal electric fields (CEF) from the surrounding ions, with exaggerated non-spherical part. If metallic and covalent bonds cause CEF of roughly equal strength (if metallic bonds are screened), oblate ∣Jb=±15/2⟩\lvert J_b = \pm15/2\rangle (prolate ∣Jb=±1/2⟩\lvert J_b = \pm1/2\rangle) is the lowest energy CEF doublet. This favours out-of-plane (in-plane) magnetization, respectively. Colour on CD isosurfaces indicates amplitude ∣E∣\lvert\mathbf{E} \rvert of the local CEF. f,g Anisotropy constants K1K_1 and K2K_2 calculated for 4f94f^{9} multiplet in DyTe3 (Methods). The abscissa describes the relative weight of CEF charges cc on Te-A and Te-B sites. Green lines bound an intermediate regime of weak K1K_1, where spin tilting and conical order are allowed.

H=∑n[J2AFMSnaSn+1a−ECDWabcos⁡(qCDWzn)SnaSnb−ECDWacsin⁡(qCDWzn)SnaSnc](1)\mathcal{H} = \sum_n \left[J_2^{\mathrm{AFM}} S_n^a S_{n+1}^a - E_{\mathrm{CDW}}^{ab} \cos\left(q_{\mathrm{CDW}} z_n\right) S_n^a S_n^b - E_{\mathrm{CDW}}^{ac} \sin\left(q_{\mathrm{CDW}} z_n\right) S_n^a S_n^c\right] \tag*{(1)}

where nn counts magnetic sites, e.g., on the upper half of the zigzag chain. The znz_n are spatial positions along the zigzag chain (cc-axis). All the coupling constants – J2AFMJ_2^{\mathrm{AFM}}, ECDWabE_{\mathrm{CDW}}^{ab}, and ECDWacE_{\mathrm{CDW}}^{ac} – are positive. The ECDWabE_{\mathrm{CDW}}^{ab} and ECDWacE_{\mathrm{CDW}}^{ac} terms are allowed by global and local mirror symmetry breaking due to the CDW, respectively. We may also introduce a Zeemann term to explain the behavior in a magnetic field and further inter-chain coupling to connect the two chains (Extended Section E1).

This 1D model naturally creates different modulation period for the aa and bcbc spin components and robustly reproduces two types of magnetic reflections, qAFM/c∗=0.5q_{\mathrm{AFM}}/c^* = 0.5 and qcyc/c∗=0.5−0.293=0.207q_{\mathrm{cyc}}/c^* = 0.5 - 0.293 = 0.207. In good consistency with experiment, Figure 4c shows that IcycI_{\mathrm{cyc}} on the order of 10 % of IAFMI_{\mathrm{AFM}} can be induced within this model. Based on scattering techniques, we find it difficult to reveal the phase-shift between CDW and the spin cycloid, and between the antiferromagnetic and incommensurate components of the magnetic order; hence, alternative (out-of-phase) locking between cycloid and antiferromagnetic component is also possible (Figure E1).

Figure E1. Two possible magnetic structures in the ground state of DyTe3\mathrm{DyTe}_{3}. Neutron scattering does not constrain the combination of cycloidal (SbS_b, ScS_c) and antiferromagnetic (AFM, SaS_a) components, which can be either in-phase (left) or out of phase (right). The AFM and cycloidal components of in-phase (out of phase) combinations are depicted, in panels b,c (in panels f, g), respectively. Only a single zigzag chain is shown, although there is antiferromagnetic coupling between subsequent zigzag chains along the bb-axis, c.f. Figure 1. d,h Magnetic texture in a single DyTe magnetic square net bilayer, corresponding to the simplified zigzag-chain picture in panels a,e. Colour on each ionic site (arrows) illustrate the bb-axis (the acac-plane) component of the magnetic moments. Dashed lines are guides to the eye, describing the square net in the upper sheet of the DyTe bilayer slab. The bright (dark) highlight marks the size of the magnetic unit cell (the size of half a magnetic unit cell, as shown in panels a-c, e-g, and in the figures of the main text). For clarity, the relative phase δ\delta between the layers is chosen so that pairs of spins are fully collinear (instead of the experimentally determined 79° phase shift, § 6.6).

Weak magneto-crystalline anisotropy

While strongly anisotropic magnetism is naively expected for dysprosium’s 6H15/2{}^{6}H_{15/2} shell, we here report a conical state with comparable mam_a, mbm_b, mcm_c in DyTe3_3. Consider the local environment of a single Dy in Figure 4d: Te-B ions form covalent bonds with the central Dy, while the point charges of Te-A are effectively screened by itinerant electrons in the conducting tellurium slab. We model the sequence of crystal electric field (CEF) states for the 4f94f^9 shell of dysprosium as a function of the effective crystal field charge cc situated on Te-A and Te-B ions (Figure E15). Figure 4e illustrates two limiting cases: When Te-A and Te-B contribute equally to the CEF, the 4f94f^9 charge cloud is compressed along the bb-direction, with ∣Jb=±15/2⟩\lvert J_b = \pm15/2\rangle dominating the ground state wavefunction, and with effective out-of-plane magnetic anisotropy for magnetic moments. Likewise, zero contribution of Te-A, i.e. highly efficient metallic screening of CEFs, favours the prolate orbital ∣Jb=±1/2⟩\lvert J_b = \pm1/2\rangle with easy-axis anisotropy.

Crystal electric field calculations and local environment of Dy in DyTe3

Figure E15. Crystal electric field calculations and local environment of Dy in DyTe3. a, Evolution of the energy gap ΔEi\Delta E_i between ground state doublet (GS, ∣ψ0±⟩\lvert\psi_0^\pm\rangle) and the first (1st ES, ∣ψ1±⟩\lvert\psi_1^\pm\rangle) or second (2nd ES, ∣ψ2±⟩\lvert\psi_2^\pm\rangle) excited state Kramers doublets. We consider a (virtual) charge transfer from the metallic Te2\mathrm{Te}_2 square net (Te-A) to the covalent bonded Tellurium (Te-B) ions, as sketched in the inset. The xx-axis is labeled by the ratio of effective point charges cc on the two types of Te ions. Within the red (grey) shaded area, the excitation gaps to the the first and second excited states are (to the first excited state is) on the order of the exchange interaction energy Eex∼2 KE_{\mathrm{ex}} \sim2\,\mathrm{K} (dashed horizontal line). Panels b-d show the composition of ∣ψ0,1,2±⟩\lvert\psi_{0,1,2}^\pm\rangle in terms of eigenstates of (b-component of) total angular momentum ∣Jb=±n/2⟩\lvert J_b = \pm n/2\rangle with n≤15n \le15, as suitable for the 4f94f^9 shell of dysprosium. The PβP_\beta are probabilities, i.e. absolute squares of amplitudes, for each contribution to the total wavefunction.

Adding exchange interactions EexE_{ex} as an effective magnetic field, the CEF Hamiltonian of a point charge model is diagonalized for the orthorhombic environment of Dy (Methods). The resulting free energy density described by two parameters K1cos⁡2(θ)+K2sin⁡2(θ)cos⁡2(ϕ)K_1 \cos^2(\theta) + K_2 \sin^2(\theta) \cos^2(\phi), where θ\theta, ϕ\phi are spherical coordinates with respect to the bb and cc crystal axes, respectively. Figure 4 f, Figure 4g testify to a transition from easy-axis to easy-plane anisotropy through a sign change of K1K_1 at intermediate charge ratio (pink line)two green lines bound the regime where easy-axis (easy-plane) anisotropy is not strong enough to prevent tilting of m\mathbf{m} along directions intermediate between bb-axis and the acac plane. Constraining EexE_{\mathrm{ex}} in agreement with TN2T_{\mathrm{N2}} and requiring easy-plane anisotropy K1>0K_1 > 0, we identify the black box in Figure 4 f, Figure 4g to capture a parameter range well consistent with experiment. Here, the model yields K2>0K_2 > 0, meaning mam_a is preferred over mcm_c.

Magnetic phase diagram and small-angle neutron scattering

We are ready, now, to consider the evolution of magnetic order in DyTe3_3 as a function of temperature and magnetic field. Figure 4a shows a contour map of the magnetic susceptibility χ\chi (Methods), where the external magnetic field is applied along the in-plane direction [101], i.e. H∥(a+c)\mathbf{H} \parallel(a+c). Heating the sample above TN1=3.6T_{\mathrm{N1}} = 3.6 K in zero field, we observe a peak splitting of qAFM\mathbf{q}_{\mathrm{AFM}}, and a concomitant shift in qcyc\mathbf{q}_{\mathrm{cyc}} that indicates the sustained coupling of the two ordering vectors, via the CDW, at elevated temperatures (Figure E13). The sharp enhancement of χc\chi_c in Figure 1a further suggests that mam_a, mbm_b survive to higher temperature than mcm_c, consistent with a putative incommensurate, fan-like order in phase II, which warrants further study.

Incommensurate and commensurate antiferromagnetic component in phase II of DyTe3

Figure E13. Incommensurate and commensurate antiferromagnetic component in phase II of DyTe3 (sample A). a,b, Line scans of magnetic intensity through the incommensurate (a) and commensurate (b) reflection on the (01L)(01L) line, with clear temperature dependence. Shaded Gaussian curves at T=3.7T = 3.7 K indicate a double-Gaussian fit in the regime of phase coexistence between phases I and II. High temperature data is multiplied by a scale factor to enhance visibility. The magnetic intensity vanishes in the incommensurate line entirely at TN2=3.85T_{\mathrm{N2}} = 3.85 K, with no indications of diffuse scattering in the thermally disordered regime. Above TN2T_{\mathrm{N2}}, we observe a weak, diffuse neutron signal along the (0,K,1/2)(0, K, 1/2) line, with rapid decay of the coherence length. c, Illustration of discommensuration-driven shift of qAFMq_{\mathrm{AFM}} in qAFM=(0,b∗,qAFM)\mathbf{q}_{\mathrm{AFM}} = (0, b^{*}, q_{\mathrm{AFM}}) for phase II, where a defect every ∼50\sim50 unit cells is introduced to release magnetoelastic strain built up between the crystal lattice and the collinear antiferromagnetic structure. c,d, Temperature dependence of qcyc\mathbf{q}_{\mathrm{cyc}} and qAFM\mathbf{q}_{\mathrm{AFM}}, respectively. The different regimes, phase I, II and the diffuse scattering regime (Diff), are highlighted by different color shadings.

To explore the regime above the critical field, confirm the coupling between qAFM\mathbf{q}_{\mathrm{AFM}} and qcyc\mathbf{q}_{\mathrm{cyc}} in phase I, and investigate the correlation of CDW and magnetic order, we carried out small angle neutron scattering (SANS) experiments in a magnetic field. Figure 5a describes the geometry of our SANS experiment and Figure 5b shows the obtained zero field (0,K,L)(0, K, L) map, while Figure 5c reduces the map into principal line cuts. The noncoplanar, helimagnetic cone texture with coupled qAFM\mathbf{q}_{\mathrm{AFM}} and qcyc\mathbf{q}_{\mathrm{cyc}} is stable up to μ0H=0.5\mu_0 H = 0.5 T for H∥c\mathbf{H} \parallel c, c.f. Figure 5f. In fact, this data further supports the existence of coupled commensurate and incommensurate order parameters in phase I, and helps to exclude a domain separation scenario (Figure E14). A pair of phase transitions to phases IV and V is visualized in Figure 5f and Figure E14. In phase V, c.f. Figure 5d and Figure 5e, which is realized when the external magnetic field exceeds μ0H=0.7\mu_0 H = 0.7 T, strong magnetic reflections appear at momentum transfer Q=(Ha∗,Kb∗,qCDW)\mathbf{Q} = (H a^{*}, K b^{*}, q_{\mathrm{CDW}}) with K=evenK = \text{even}, demonstrating direct coupling between incommensurate magnetism and the CDW in absence of the antiferromagnetic order parameter. Although we cannot provide a magnetic structure model based on the available data, the comparison of intensities at Miller indices K=0,2K = 0, 2 suggests the presence of both mam_a and mbm_b spin components (inset of Figure 5e).

Image not extracted

Figure 5. Coupling of magnetic order and charge-density wave in DyTe3, tracked by small-angle neutron scattering (SANS) in a magnetic field along the cc-axis at T=2T = 2 K. a, Experimental geometry of SANS. Orange plate, blue rectangle, kin\mathbf{k}_{\mathrm{in}}, kout\mathbf{k}_{\mathrm{out}}, and Q\mathbf{Q} are the sample, the area detector – shifted from the center axis to catch reflections with larger ∣Q∣\lvert\mathbf{Q}\rvert –, incoming / outgoing neutron beam wavevector, and the momentum transfer of elastic scattering, respectively. A wedge of three-dimensional Q\mathbf{Q} space is covered by rotating the sample in steps (angle ω\omega, Methods). b-e, Scattering intensity in (0KL)(0KL) cuts of momentum space (upper) and linecuts along (00L)(00L) (lower), for intensity integrated along the KK-direction within the orange and red dashed boxes. The magnetic field induces a transition to ferroic stacking, K=0K = 0. f, Integrated intensity of various reflections as a function of magnetic field, covering three magnetic phases. Inset of (e): Experimental ratio γ=I(0,−2,−0.293)/I(0,0,−0.293)\gamma= I(0, -2, -0.293)/I(0, 0, -0.293) in Phase V, compared to simulations for fan-like orders (spin plane indicated, green and grey dashed lines) and a helimagnetic (longitudinal-cone, red dashed line) order. The experiments are consistent with a slightly distorted, conical state of the same period as the underlying charge order.

Figure E14. Coupling of commensurate and incommensurate magnetic order in DyTe₃, tracked by small-angle neutron scattering (SANS) in a magnetic field along the c-axis, for sample E. a, Integrated intensity of commensurate (q=qAFM\mathbf{q} = \mathbf{q}_{\mathrm{AFM}}) and incommensurate (q=qcyc\mathbf{q} = \mathbf{q}_{\mathrm{cyc}}) reflections of type (0,−1,−L)(0, -1, -L) as a function of magnetic field and at a temperature of T=2T = 2 K. The observed intensity of both drops simultaneously at around μ0H=0.4\mu_{0}H = 0.4 T, with a fixed intensity ratio I(qAFM)/I(qcyc)∼1.5I(\mathbf{q}_{\mathrm{AFM}})/I(\mathbf{q}_{\mathrm{cyc}}) \sim1.5, as indicated by a grey dashed line in panel b. The error bars correspond to Poisson counting errors of the integrated neutron scattering intensity.

Discussion

As compared to transition metal dichalcogenides, where the magnetic ion is buried inside a rather symmetric block layer [34, 35], RTe3\mathrm{RTe}_{3} harbors more complex structural features, with magnetic ions at the boundary between metallic and covalently bonded blocks. This mixed covalent / metallic environment for the magnetic ion is key to realizing the present scenario: it facilitates coupling between magnetic ions and a charge density wave (CDW) on the tellurium square net, and – at the same time – generates unconventional magnetocrystalline anisotropy. In fact, the present charge-transfer phenomenology is partially inspired by work on thin films of magnetic metals on insulating substrates [36], on electric field control of magnetocrystalline anisotropy [37], and on the behaviour of magnetic materials when charge transfer is induced by oxidation at the surface [38].

Symmetry breaking with cycloid / spiral magnetic order of fixed helicity is rather widely observed in zigzag chain magnets (Extended Section E4 A), but the present combination of antiferromagnetic and cycloidal components is unique. For example, Mn2GeO4\mathrm{Mn}_{2}\mathrm{GeO}_{4} has cones arrayed on one-dimensional chains, with uniform cone direction along the chain [39]. In another metallic system, EuIn2As2\mathrm{EuIn}_{2}\mathrm{As}_{2}, jumps in the rotation sense of a helimagnetic texture have recently been identified, with a short magnetic period [40]. In contrast, the rotation of moments in DyTe3\mathrm{DyTe}_{3} proceeds in nearly parallel pairs, without abrupt jumps in the cycloidal component of the texture. We expect helimagnetic orders of the type observed here to be common in layered materials, and especially in rare earth tellurides and selenides. Here, rich magnetic phase diagrams have been generally observed [41, 42] and could be amenable to modeling by CDW-induced terms as in Eq. (1).

This complex magnetic order, its relationship to a strain-controllable CDW [33], and its (likely) rich excitation spectrum certainly warrant further research. For example, the CDW’s gapped collective excitation, termed Higgs mode, shows a magnetic character in RTe3\mathrm{RTe}_{3} as observed via Raman scattering experiments [15], and its evolution below TN1T_{\mathrm{N1}} may provide insights on both the origin of magnetic order and the nature of the CDW in DyTe3DyTe_{3}.

Furthermore, the lowest-energy, Goldstone mode of a typical helimagnet corresponds to a spatial shift of the magnetic texture, termed phason excitation [43]. In DyTe3DyTe_{3}, the magnetic and CDW phasons [44] are expected to be closely intertwined, as evident from the robust qcyc(T)q_{\mathrm{cyc}}(T) in phase I, and its jump – by the same amount as qAFMq_{\mathrm{AFM}} – in phase II (Figure E13). Such locking between low-energy modes may have implications for dynamic responses, further enriching the spectrum of elementary excitations in RTe3RTe_{3}.

Concluding remarks

An important open question is the stability of helimagnetism in few-layer devices of DyTe3DyTe_{3}, where the cleavage plane, as well as the center of structural inversion, are situated between tellurium bilayers. As a fundamental building block of the structure, we consider a DyTe slab sandwiched by Te square nets – that is half a unit cell in Figure 1 a. Being screened from top and bottom by tellurium layers, we expect no qualitative change of the local crystal field environment of Dy in the few-layer limit. However, the absence of an inversion center for odd numbers of layers, and its presence for even numbers of layers, may have a profound effect on magnetic ordering and the presence or absence of (right- or left-handed) helicity domains in the sample, considering the presence or absence of Dzyaloshinskii-Moriya interactions [45].

Most appealingly, DyTe3DyTe_{3} is a potential platform for spin-Moiré engineering in solids, where complex magnetic textures can be designed by combining and twisting two or more helimagnetic sheets. Here, a plethora of noncoplanar spin textures can be engineered at will [6, 46], while highly conducting tellurium square net channels may serve as a test bed for of emergent electromagnetism in a tightly controlled setting [47, 48].

References

Methods

Sample preparation and characterization

Single crystals are grown from tellurium self-flux following the recipe in Ref. [42]: We set elemental Dy and Te at a ratio of 1 : 21.65 in an alumina crucible, which in turn is sealed in a quartz tube in high vacuum. The raw materials are heated to 450 °C for 36 hours and then to 780 °C in 96 hours, where the melt remained for 48 hours, followed by cooling to 450 °C at a rate of 1.375 °C/hour. The final product is centrifuged after renewed heating to 500° Celsius, so that plate-shaped single crystals of typical dimensions 5×5×1 mm35 \times5 \times1\ \mathrm{mm}^{3} are obtained. The face of each plate is perpendicular to the bb-axis of DyTe3\mathrm{DyTe}_{3}’s orthorhombic unit cell, and facet edges tend to be parallel to either aa or cc. The existence of impurity phases above 1% volume fraction is ruled out by single-crystal x-ray diffraction on cleaved surfaces in a Rigaku SmartLab X-ray powder diffractometer. The experiment yields lattice constants of a=4.27(2)a = 4.27(2) Å, b=25.433(1)b = 25.433(1) Å, and c=4.27(2)c = 4.27(2) Å at room temperature, in good agreement with previous work [18]. We found it challenging to obtain high-quality powder x-ray data from crushed single crystals, which include traces of Te flux on their surface and form thin flakes, even when thoroughly ground in a mortar. We also verified the stoichiometric chemical composition of our crystals by energy-dispersive x-ray spectroscopy (EDX). Cleaved single crystals have a reddish-brown surface; but even in vacuum, the colour of the surface changes to silver-metallic, and then to black, after two weeks or so. A red hue can be recovered by renewed surface cleaving.

Magnetization measurements and crystal alignment

We use a commercial magnetometer with T=2T = 2 K base temperature and a maximum magnetic field of 7 T (MPMS, Quantum Design, USA). The measurement is carried out using a rectangular-shaped single crystal of mass m=1.22m = 1.22 mg, with carefully aligned edges along the aa and cc crystal axes. By means of a single crystal diffractometer (Malvern Panalytical Empyrean, Netherlands), we confirm the extinction rule h+k=evenh+k = \mathrm{even} in space group Cmcm\mathit{Cmcm}. It is difficult to distinguish aa and cc axes in this orthorhombic, yet nearly tetragonal structure by eye or with the help of the Laue diffractogram. Temperature dependent susceptibility χ(T)\chi(T) is measured in a DC magnetometer with 1000 Oe applied field; there is no observable difference between field-cooled and zero-field cooled magnetization traces. A demagnetization correction is carried out according to the standard expression Hint=Hext−NM\mathbf{H}_{\mathrm{int}} = \mathbf{H}_{\mathrm{ext}} - N\mathbf{M}, where Hext\mathbf{H}_{\mathrm{ext}}, M\mathbf{M}, and NN are the externally applied magnetic field, the bulk magnetization, and the dimensionless demagnetization factor. The latter is calculated by approximating the crystal as an oblate ellipsoid [49]. For the H−TH-T phase diagram in Figure 4, the [101][101] direction is aligned within ±3∘\pm3^{\circ} and bulk magnetization is measured in discrete field steps, for selected temperatures. Figure 4 shows data for decreasing magnetic field ∂H/∂t<0\partial H/\partial t < 0. Note that hysteresis occurs at all phase transitions shown in Figure 4a, indicating their first-order nature. The magnetic anisotropy energy is expressed as E/Vuc=K0+K1cos⁡2(θ)+O[sin⁡4(θ)]E/V_{\mathrm{uc}} = K_0 + K_1 \cos^2(\theta) + \mathcal{O}[\sin^4(\theta)], where θ\theta is the angle between M\mathbf{M} and the bb-axis. Utilizing the free energy expression F=F0+∑αai(T)Mα2+O(M4)F = F_0 + \sum_\alpha a_i(T)M_\alpha^2 + \mathcal{O}(\mathbf{M}^4) and the Curie-Weiss law χα=C/(T−ΘCWα)\chi_\alpha= C/(T-\Theta_{\mathrm{CW}}^\alpha), where ΘCWα\Theta_{\mathrm{CW}}^\alpha is the Curie-Weiss temperature along the α∈(a,b,c)\alpha\in(a,b,c) direction and C=2.077(1) KC = 2.077(1)\ \mathrm{K} is the Curie constant of DyTe3_3, we obtain K1=(μ0/2)(M/Vuc)2(ΔΘCW/C)=22(1) kJ m−3K_1 = (\mu_0/2)(M/V_{\mathrm{uc}})^2(\Delta\Theta_{\mathrm{CW}}/C) = 22(1)\ \mathrm{kJ\,m^{-3}}, where ΔΘCW=ΘCWac−ΘCWb=−2.0(1) K\Delta\Theta_{\mathrm{CW}} = \Theta_{\mathrm{CW}}^{ac} - \Theta_{\mathrm{CW}}^b = -2.0(1)\ \mathrm{K} is the difference between the Curie-Weiss temperatures in the acac-plane and along the bb-axis (c.f. Figure 1d, inset). Specific heat was recorded using a relaxation technique in a Quantum Design PPMS cryostat, in zero magnetic field. For specific heat anomalies in applied magnetic field, we employed the AC calorimetry technique in a custom-built setup. Anisotropy of the resistivity, as in Figure 2, was recorded on exfoliated flakes of thickness ∼100 μm\sim100\ \mu\mathrm{m} using the Montgomery technique. Electric contacts are made with Ag paste (Dupont) and deteriorate with time. To maintain excellent contact resistance ∼1 Ω\sim1\ \Omega, it is crucial to immediately cool the contacted crystal in vacuum, after depositing the silver paste. The sample and contact quality is robust at low temperatures for at least two weeks.

Elastic neutron scattering

We performed unpolarized and polarized neutron scattering experiments using the POlarized Neutron Triple-Axis spectrometer (PONTA) installed at the 5G beam hole of the Japan Research Reactor 3 (JRR-3). Two single crystals of DyTe3_3 (Sample A and B) are cut into rectangular shapes with dimension 3.6×2.7×0.9 mm3.6 \times2.7 \times0.9\ \mathrm{mm} and 1.7×1.9×0.8 mm1.7 \times1.9 \times0.8\ \mathrm{mm}, respectively. For both samples, the widest surface is normal to the bb-axis, and the sides are parallel to the aa or cc-axis. Each sample is set in an aluminum cell, which is sealed with 4He^4\mathrm{He} gas for thermal exchange. We employed a 4He^4\mathrm{He} closed-cycle refrigerator with base temperature of 2.2 K2.2\ \mathrm{K}, and measured intensities on the (0,K,L)(0,K,L) horizontal scattering plane. Using a PG (002) monochromator, the energy of the incoming neutron beam is set to Ei=14.7 meVE_i = 14.7\ \mathrm{meV} (30.5 meV) for unpolarized measurements of sample A (sample B). For the unpolarized measurements, the spectrometer is operated in two-axis mode with horizontal beam collimation of open-80′80^{\prime}-80′80^{\prime}. In both unpolarized and polarized experiments, sapphire and pyrolytic graphite (PG) filters are installed between the monochromator and the sample, to suppress higher-order reflections from the monochromator to less than 0.5%. The observed integrated intensities are converted to structure factors after applying the Lorentz factor and absorption corrections.

For sample B, we measured nuclear and magnetic Bragg reflections at 2.2 K by θ−2θ\theta- 2\theta scans. For the scattering profiles showing a well-defined Gaussian-shape peak, we estimated the background from both ends of the profile. For the magnetic reflections located near the powder diffraction lines of the Al sample holder, we carried out background scans at 10 K, and subtracted the intensities from those measured at 2.2 K. We also measured the background data at 10 K for relatively weak commensurate magnetic reflections in the QQ-range of ∣Q∣>4.0 A˚−1|Q| > 4.0\ \text{\AA}^{-1}, to check for possible λ/2\lambda/2 contamination from the nuclear reflections. As for the absorption correction, we calculated the scattering path length ll inside the sample, based on the dimensions of Sample B and on the incident and scattered directions of the neutrons. The neutron transmission is given by exp⁡(−μl)\exp(-\mu l), where μ\mu is the linear absorption coefficient. Taking into account the incident energy and the absorption and incoherent scattering cross-sections of DyTe3_3, μ\mu is calculated to be 8.392 cm−18.392\ \mathrm{cm}^{-1}.

The diffraction profiles and integrated intensities shown in Figure 2, Figure 3 were measured using Sample A. Contrary to integrated intensities in the case of refinement, the temperature dependences in Figure 2c, Figure E13 are obtained from LL-scans of magnetic scattering. The calculation of the scattering intensity in Figure 2f, which includes the third harmonic reflection, takes into account instrumental resolution broadening (Figure E10), anharmonicity of the cycloidal magnetic structure component, and the presence of two magnetic domains (Extended Section E4).

Calibration of instrument resolution at 5G-PONTA

Figure E10. Calibration of instrument resolution at 5G-PONTA at the JRR-3 neutron reactor source, for sample A. Each pair of (red, blue) data points corresponds to the full width at half maximum (FWHM) of a nuclear lattice reflection. The shape of the resolution ellipsoid in the b∗b^{*}–c∗c^{*} plane is defined by the width of an ω\omega-scan (wTw_{\mathrm{T}}, transverse width) and the width of a ω−2θ\omega- 2\theta scan (wLw_{\mathrm{L}}, longitudinal width), respectively. These parameters depend on the sample shape, crystal quality, the momentum transfer ∣Q∣|\mathbf{Q}|, and the performance of the instrument. Blue and red dashed lines are a linear fit, and a second order polynomial fit to the data, respectively. These fits can be used for estimation of the instrument resolution at arbitrary positions in the bcbc scattering plane, e.g. in Figure 4 c. Error bars correspond to statistical uncertainties of Gaussian fits to the nuclear reflections.

Sample A is also used for polarized neutron scattering, in which the spectrometer is operated in the triple-axis mode with horizontal beam collimation of open-80′80^{\prime}-80′80^{\prime}-open. A polarized neutron beam with Ei=13.7 meVE_i = 13.7\ \mathrm{meV} is obtained by a Heusler (111) crystal monochromator. The spin direction of incident neutrons is set to be perpendicular to the scattering plane. We thus applied weak vertical magnetic fields of approximately 5 mT throughout the beam path by guide magnets and a Helmholtz coil. We used a Mezei-type π\pi spin flipper placed between monochromator and sample, and employed a Heusler (111) crystal analyzer to select the energy and spin states of scattered neutrons, separating spin-flip (SF) and non-spin flip (NSF) intensities. The spin polarization of the incident neutron beam (P0P_0) is 0.823, as measured using the (002) nuclear Bragg reflection of the sample.

Small angle neutron scattering in magnetic field

SANS measurements were performed using the SANS-I instrument at Paul Scherrer Institute (PSI), Switzerland. A bulk single crystal of DyTe3_3 (Sample E, m=73.4m = 73.4 mg) was carefully aligned (c.f. crystal alignment methods) and installed into a 1.8 T horizontal-field cryomagnet so that the aa-axis is vertical, and the incident neutron beam is in the bcbc-plane. The magnetic field is applied parallel to the crystal cc-axis, as shown in Figure 5a. The incident neutron beam with λ=3.1\lambda= 3.1 Å wavelength (15% Δλ/λ\Delta\lambda/\lambda) is collimated over a distance of 4.5 m before the sample, and the scattered neutrons are detected by a 1 m2^2 two-dimensional multidetector (pixel size 7.5 mm x 7.5 mm) placed 1.7 m behind the sample. To cover a broader qq-space up to qAFMq_{\mathrm{AFM}} along the (01LL) direction, the detector was also translated 0.45 m in the horizontal plane. For all SANS data, background signals from the sample and the instrument are subtracted using the data of the nonmagnetic state at T=10T = 10 K and μ0H=0\mu_0 H = 0 T. The field-dependent SANS measurements are performed during a field-increasing process, after an initial zero-field cool to the base temperature of 2 K. For each measurement, rocking scans were performed, i.e. the cryomagnet is rotated together with the sample around the vertical crystal aa-axes (rocking angle ω\omega) in a range from −102∘-102^\circ to 55∘55^\circ and steps of 2∘2^\circ (−38∘≤ω≤28∘-38^\circ\leq\omega\leq28^\circ) and 1∘1^\circ (else). Here ω=0∘\omega= 0^\circ is carefully aligned and corresponds to the configuration where the beam kin\mathbf{k}_{\mathrm{in}} is parallel to the crystallographic bb-axis. The SANS maps shown in this paper are obtained by performing a 2D cut of the volume of reciprocal space measured through cumulative detector measurements taken at each angle of the rocking scan. For the SANS maps shown in Figure 5, the integration width along the out-of-plane (H00H00) direction is ±0.15\pm0.15 reciprocal lattice units. The line cuts along (0KfixL)(0K_{\mathrm{fix}}L) shown in Figure 5c, Figure 5e are extracted by integrating over a region of ±0.2\pm0.2 reciprocal lattice units in the (0K0)(0K0) direction. Peak positions and integrated intensities are calculated using those linecuts and a multi-peak fitting.

Crystal electric field calculations

We use the software package PyCrystalField [50] for the calculation of crystal electric field energies via the point charge model in the limit of strong spin-orbit interactions. The calculation is based on published fractional coordinates of Dy and Te ions within the crystallographic unit cell [18, 22], with a ∼0.15%\sim0.15\% tensile strain along the aa-axis, lifting tetragonal symmetry and yielding finite K2K_2. In Figure E15, we vary the effective crystal electric field originating from Te-A (on the Te2_2 slab) by changing its point charge, while keeping the total charge in the environment of Dy unchanged. An unperturbed, diagonal Hamiltonian matrix is constructed from the energies in Figure E15, and the operator of total angular momentum J=L+S\mathbf{J} = \mathbf{L} + \mathbf{S} is also expressed in the basis of these CEF eigenstates. Adding an effective exchange field EexJαE_{\mathrm{ex}}J_\alpha (α=a,b,c\alpha= a,b,c are vector components), the total Hamiltonian is diagonalized and the expectation value of JaJ_a, JbJ_b, JcJ_c is evaluated in the respective ground state. The anisotropy constants are approximated, as

K1∝1⟨Jb⟩b2−1⟨Ja⟩a2(2)K_1 \propto\frac{1}{\langle J_b\rangle_b^2} - \frac{1}{\langle J_a\rangle_a^2} \tag*{(2)}
K2∝1⟨Jc⟩c2−1⟨Ja⟩a2(3)K_2 \propto\frac{1}{\langle J_c\rangle_c^2} - \frac{1}{\langle J_a\rangle_a^2} \tag*{(3)}

so that K1<0K_1 < 0 for easy-axis anisotropy along the bb-axis and K2>0K_2 > 0 if aa-axis orientation is energetically preferred over the cc-axis. Here, ⟨Jα⟩α\langle J_\alpha\rangle_\alpha is shorthand for ⟨ψ0,α∣Jα∣ψ0,α⟩\langle\psi_{0,\alpha}\lvert J_\alpha\rvert\psi_{0,\alpha}\rangle, where ∣ψ0,α⟩\lvert\psi_{0,\alpha}\rangle is the ground state of the total Hamiltonian when an exchange field of magnitude EexE_{\mathrm{ex}} is applied along the α\alpha-direction.

The anisotropic part of the charge density is exaggerated 20×20\times in Figure 4 according to the expression 20⋅(R−R0)+R020\cdot(R-R_0)+R_0, where R0R_0 corresponds to 10 Bohr radii. More details are given in Figure E15, Figure E16.

Six panels showing mixing of crystal field states in DyTe3 by exchange interactions

Figure E16. Mixing of crystal field states in DyTe3 by exchange interactions. a-c, The maximal and minimal eigenvalues of J^a\hat{J}_a and d-f of J^b\hat{J}_b calculated for the crystal field doublets ∣ψ0,1,2±⟩\lvert\psi_{0,1,2}^{\pm}\rangle. The charge ratio (xx-coordinate) describes the ratio of effective charges assigned to Te-A and Te-B tellurium ions corresponding to metallic and ionic bonds around Dy, respectively. A matrix representation of the operators J^a\hat{J}_a and J^b\hat{J}_b is calculated on each respective subspace spanned by a Kramers pair of states, and subsequently diagonalized. There is a transition from dominant ⟨ψ0∣J^b∣ψ0⟩=±1/2\langle\psi_0\lvert\hat{J}_b\lvert\psi_0\rangle=\pm1/2 (easy-plane) to ±15/2\pm15/2 (easy-axis) for the ground state ∣ψ0⟩\lvert\psi_0\rangle, when reducing the effective crystal electric field charge for Te-A on metallic bonds. The first excited state behaves similarly, while the second excited state generally covers a broader range of J^b\hat{J}_b eigenvalues. Red and grey shaded areas are defined as in Figure E15. In the red region, the character of the ground state (GS, ∣ψ0±⟩\lvert\psi_0^{\pm}\rangle, left column) changes from predominant ∣Jb=±15/2⟩\lvert J_b=\pm15/2\rangle to mainly ∣Jb=±1/2⟩\lvert J_b=\pm1/2\rangle character. It is not possible to generate sizable in-plane (bb-axis) magnetic moment using any linear combination of ∣ψ0+⟩\lvert\psi_0^+\rangle, ∣ψ0−⟩\lvert\psi_0^-\rangle at relative weight of CEF charges qq on Te-A and Te-B sites >0.5>0.5 (<0.4<0.4).

Spin model calculations

A model Hamiltonian, Eq. (1), is introduced from the viewpoint of symmetry in § 6.1 of Extended Data, to explain the essential experimental results. Here, magnetic frustration is lifted and the separation of spin components by modulation (q\mathbf{q})-vectors is explained naturally by the off-diagonal ECDWabE_{\mathrm{CDW}}^{ab} and ECDWacE_{\mathrm{CDW}}^{ac} terms. An analytic solution is obtained in Fourier space, and variational calculations are carried out based on the spin ansatz and ignoring higher harmonics, for simplicity,

(Sna,Snb,Snc)=((−1)nm,1−m2cos⁡[(π+qCDW)zn],1−m2sin⁡[(π+qCDW)zn]),(4)\left(S_n^a,S_n^b,S_n^c\right)=\left((-1)^n m,\sqrt{1-m^2}\cos\left[(\pi+q_{\mathrm{CDW}})z_n\right],\sqrt{1-m^2}\sin\left[(\pi+q_{\mathrm{CDW}})z_n\right]\right), \tag*{(4)}

In these terms, the energy is given by

E[m]=−J2AFMm2−ECDWab+ECDWac2m1−m2(5)E[m]=-J_2^{\mathrm{AFM}}m^2-\frac{E_{\mathrm{CDW}}^{ab}+E_{\mathrm{CDW}}^{ac}}{2}m\sqrt{1-m^2} \tag*{(5)}

and easily optimized at m=m∗m=m^* satisfying δE[m∗]=0\delta E[m^*]=0. The optimal antiferromagnetic moment m∗m^* gives the squared intensities IAFM=∣Sa(q=π)∣2=m2I_{\mathrm{AFM}}=\left|S^a(q=\pi)\right|^2=m^2 and Icyc=∣Sb(q=π+qCDW)∣2+∣Sc(q=π+qCDW)∣2=(1−m2)/2I_{\mathrm{cyc}}=\left|S^b(q=\pi+q_{\mathrm{CDW}})\right|^2+\left|S^c(q=\pi+q_{\mathrm{CDW}})\right|^2=(1-m^2)/2, depicted in Figure 4.

Acknowledgments

We thank M. Nakano for permission to use his single-crystal x-ray diffractometer, and for support during the measurement. Moreover, we thank S. Kitou for initial advice on crystal field calculations, M. Kriener for support with experiments on magnetization and calorimetry, and Y. Kato for critical advice on the theoretical spin model. We thank P.M. Neves for support with the analysis of the wide-rocking angle SANS diffraction data. This work is based partly on experiments performed at the Swiss spallation neutron source SINQ, Paul Scherrer Institute, Villigen, Switzerland. This work was supported by JSPS KAKENHI Grant Nos. JP22H04463, JP22F22742, JP22K13998, JP23H01119, JP23KJ0557 and JP22K20348 as well as JST CREST Grant Number JPMJCR1874 and JPMJCR20T1 (Japan), and JST FOREST JPMJFR2238 (Japan). M.M.H. was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – project number 518238332. The authors are grateful for support by the Fujimori Science and Technology Foundation, New Materials and Information Foundation, Murata Science Foundation, Mizuho Foundation for the Promotion of Sciences, Yamada Science Foundation, Hattori Hokokai Foundation, Iketani Science and Technology Foundation, Mazda Foundation, Casio Science Promotion Foundation, Inamori Foundation, Marubun Exchange Grant, and Kenjiro Takayanagi Foundation.

Data availability

The data supporting the findings of this study are available from the authors upon reasonable request.

Author contributions

Sh.A., M.H., and Y.O. synthesized and characterized the single-crystals. S.E., M.H., and Sh.A. carried out calorimetry and magnetic measurements, while T.N., Se.A., Sh.A., Ri.Y., S.E. and M.H. carried out and analyzed elastic neutron scattering measurements, with extensive guidance from T.-h.A. S.E., J.S.W. and M.H. carried out and analyzed small angle neutron scattering measurements. Electric transport measurements were carried out by S.E. and M.H., and theoretical modeling of crystal fields was conducted by S.E. under the guidance of T.-h.A. S.G., S.O., and Ry.Y. modeled the magnetic structure. M.H., M.M.H., and S.E. wrote the manuscript with contributions and comments from all co-authors.

Competing interests

The authors declare no competing interests.

Spin Hamiltonian

Model Hamiltonian. To describe the ground state and the field-induced transition in DyTe3_3 for B∥c\mathbf{B} \parallel c, consider a one-dimensional (1D) chain, where each unit cell contains a single magnetic site, of index nn, in the paramagnetic state. For this model, the lattice constant is set to c=1c=1, the reciprocal lattice constant is c∗=2πc^{*}=2\pi, the wavevector qq is dimensionless, and the CDW wavenumber (for DyTe3DyTe_{3}) is qCDW=2π⋅0.293q_{\mathrm{CDW}}=2\pi\cdot0.293 [18]. As compared to the effective zigzag chain in DyTe3DyTe_{3}, the number of magnetic sites is halved, i.e. the lower part of the chain is omitted. Hence, udud order for the component SaS^{a} of the local, quasi-classical spin S\mathbf{S} is equivalent to uudduudd order on the zigzag chain.

We start with a real-space ansatz comprising antiferromagnetic exchange, a spatially modulated on-site coupling induced by the lattice distortion from the CDW, and a Zeeman term for magnetic field applied along the chain axis (cc-axis),

H=∑n[J2SnaSn+1a−ECDWabcos⁡(qCDWzn)SnaSnb−ECDWacsin⁡(qCDWzn)SnaSnc+BSnc](6)\mathcal{H}=\sum_{n}\left[J_{2}S_{n}^{a}S_{n+1}^{a}-E_{\mathrm{CDW}}^{ab}\cos\left(q_{\mathrm{CDW}}z_{n}\right)S_{n}^{a}S_{n}^{b}-E_{\mathrm{CDW}}^{ac}\sin\left(q_{\mathrm{CDW}}z_{n}\right)S_{n}^{a}S_{n}^{c}+BS_{n}^{c}\right] \tag*{(6)}

where nn labels magnetic moments on a single layer of a single zigzag chain. Here, nn, n+1n+1 represent nearest neighbours in the crystal lattice of DyTe3DyTe_{3}, so that their coupling J2J_{2} can be expected to be stronger than the coupling J1J_{1} between the sheets (see next section).

The oscillating off-diagonal terms, ECDWabcos⁡(qCDWzn)E_{\mathrm{CDW}}^{ab}\cos\left(q_{\mathrm{CDW}}z_{n}\right) and ECDWacsin⁡(qCDWzn)E_{\mathrm{CDW}}^{ac}\sin\left(q_{\mathrm{CDW}}z_{n}\right) are permitted since the global Mb\mathcal{M}_{b} and the local Mc\mathcal{M}_{c}-mirror are broken by the lattice distortion due to the CDW, respectively. Moving to Fourier space according to the conventions Sn=(1/N)∑qSqexp⁡(iqzn)\mathbf{S}_{n}=(1/\sqrt{N})\sum_{q}\mathbf{S}_{q}\exp(iqz_{n}) and ∑nexp⁡(iqzn)=Nδ(q)\sum_{n}\exp(iqz_{n})=N\delta(q), where NN is the number of sites on the chain, we have

H=∑q[J2cos⁡(q)SqaS−qa−ECDWab2Sqa(S−q+qCDWb+S−q−qCDWb)−ECDWac2iSqa(S−q+qCDWc−S−q−qCDWc)]+BNSq=0c(7)\mathcal{H}=\sum_{q}\left[J_{2}\cos(q)S_{q}^{a}S_{-q}^{a}-\frac{E_{\mathrm{CDW}}^{ab}}{2}S_{q}^{a}\left(S_{-q+q_{\mathrm{CDW}}}^{b}+S_{-q-q_{\mathrm{CDW}}}^{b}\right)-\frac{E_{\mathrm{CDW}}^{ac}}{2i}S_{q}^{a}\left(S_{-q+q_{\mathrm{CDW}}}^{c}-S_{-q-q_{\mathrm{CDW}}}^{c}\right)\right]+\frac{B}{\sqrt{N}}S_{q=0}^{c} \tag*{(7)}

For a Heisenberg Hamiltonian without off-diagonal terms, the Luttinger-Tisza rule dictates the choice of a single, optimal qq for the long-range ordered state; yet here, given the off-diagonal coupling of spin components, we naturally select different wavenumbers for different spin components, explaining the clear separation of spin components by qq-vector observed experimentally in Figure 2 of the main text. Hereafter, we set ECDWab=ECDWac=ECDWE_{\mathrm{CDW}}^{ab}=E_{\mathrm{CDW}}^{ac}=E_{\mathrm{CDW}} for simplicity.

If J2>0J_{2}>0, the minimum of J2cos⁡(q)J_{2}\cos(q) is at q=πq=\pi and SπaS_{\pi}^{a} is realized. If further ECDW>0E_{\mathrm{CDW}}>0, SqcS_{q}^{c} is directly locked to the dominant SqaS_{q}^{a} and

Sqb∝δ(q=π+qCDW)+δ(q=π−qCDW)(8)S_{q}^{b}\propto\delta(q=\pi+q_{\mathrm{CDW}})+\delta(q=\pi-q_{\mathrm{CDW}}) \tag*{(8)}
Sqc∝δ(q=π+qCDW)−δ(q=π−qCDW)(9)S_{q}^{c}\propto\delta(q=\pi+q_{\mathrm{CDW}})-\delta(q=\pi-q_{\mathrm{CDW}}) \tag*{(9)}

realizing the coupled wavevectors that are experimentally observed in phase I of DyTe3_3. Application of a magnetic field BB larger than J2J_2 enforces Sqc∝δ(q=0)S^c_q \propto\delta(q=0) and the cross-terms induce Sqa∝δ(q=±qCDW)S^a_q \propto\delta(q=\pm q_{\mathrm{CDW}}) and the second harmonics Sqb∝δ(q=±2qCDW)S^b_q \propto\delta(q=\pm2q_{\mathrm{CDW}}) demonstrated experimentally in Figure 5. Indeed, the Zeeman energy per magnetic moment of mDy=10 μBm_{\mathrm{Dy}}=10\,\mu_{\mathrm{B}} (Bohr magneton) at the critical field Bc=0.5 TB_{\mathrm{c}}=0.5\,\mathrm{T} in Figure 5 is roughly 3.6 K, very close to the value of the Néel temperature TNT_{\mathrm{N}}.

If, in contrast, ECDW<0E_{\mathrm{CDW}}<0, there is no reason why wavenumbers such as π±qCDW\pi\pm q_{\mathrm{CDW}} should appear in the ground state. As compared to anisotropic exchange interactions, for example of the type SncSn+1a+SnaSn+1cS^c_n S^a_{n+1}+S^a_n S^c_{n+1}, the present ECDWE_{\mathrm{CDW}} terms cannot induce spontaneous magnetic order by themselves, but rather they create a ’parasitic’ spin modulation – driven by the charge-density wave of RTe3R\mathrm{Te}_3, R=R= rare earth – on the back of either AFM order below BcB_{\mathrm{c}} or of the field-polarized moment above BcB_{\mathrm{c}}.

Additional Heisenberg coupling on zigzag chain. We introduce an additional J1J_1 between neighbouring sites on the zigzag chain as

H′′=J1∑n(Sn1⋅Sn2+Sn2⋅Sn+1,1)(10)\mathcal{H}^{\prime\prime}=J_1\sum_n\left(\mathbf{S}_{n1}\cdot\mathbf{S}_{n2}+\mathbf{S}_{n2}\cdot\mathbf{S}_{n+1,1}\right) \tag*{(10)}

which can be separated into three independent equations Ha′′\mathcal{H}^{\prime\prime}_a, Hb′′\mathcal{H}^{\prime\prime}_b, Hc′′\mathcal{H}^{\prime\prime}_c – one for each spin component. Note that Sn1\mathbf{S}_{n1}, Sn+1,1\mathbf{S}_{n+1,1} are not nearest neighbours in the lattice of DyTe3_3, so that their coupling J1J_1 is expected to be weaker than the dominant antiferromagnetic J2J_2 in Eq. (6). We use the trial functions

Snαa=Sˉacos⁡(2π⋅12⋅znα+φαAFM)(11)S^a_{n\alpha}=\bar{S}_a\cos\left(2\pi\cdot\frac{1}{2}\cdot z_{n\alpha}+\varphi^{\mathrm{AFM}}_\alpha\right) \tag*{(11)}
Snαb=Sˉbsin⁡(hα⋅qcyc⋅znα+φα)(12)S^b_{n\alpha}=\bar{S}_b\sin\left(h_\alpha\cdot q_{\mathrm{cyc}}\cdot z_{n\alpha}+\varphi_\alpha\right) \tag*{(12)}
Snαc=Sˉccos⁡(hα⋅qcyc⋅znα+φα)(13)S^c_{n\alpha}=\bar{S}_c\cos\left(h_\alpha\cdot q_{\mathrm{cyc}}\cdot z_{n\alpha}+\varphi_\alpha\right) \tag*{(13)}

with a helicity parameter hαh_\alpha indicating the propagation direction of the texture, the incommensurate wavevector qcyc=2πLcycq_{\mathrm{cyc}}=2\pi L_{\mathrm{cyc}} in reciprocal lattice units (r.l.u.), as well as the phases φα\varphi_\alpha and φαAFM\varphi^{\mathrm{AFM}}_\alpha. The latter can only vary in steps of π\pi, if we assume a fixed length of the

magnetic moment. Based on this,

Hb′′=J1S‾b2cos⁡(h1qcyc/2)∑n[cos⁡((h2−h1)qcyc(n+1/2)+φ2−φ1)−cos⁡((h2+h1)qcyc(n+1/2)+φ2+φ1)](14)\mathcal{H}^{\prime\prime}_{b}=J_{1}\overline{S}_{b}^{2}\cos(h_{1}q_{\mathrm{cyc}}/2)\sum_{n}\left[\cos\left((h_{2}-h_{1})q_{\mathrm{cyc}}(n+1/2)+\varphi_{2}-\varphi_{1}\right)-\cos\left((h_{2}+h_{1})q_{\mathrm{cyc}}(n+1/2)+\varphi_{2}+\varphi_{1}\right)\right] \tag*{(14)}
Hc′′=J1S‾c2cos⁡(h1qcyc/2)∑n[cos⁡((h2−h1)qcyc(n+1/2)+φ2−φ1)+cos⁡((h2+h1)qcyc(n+1/2)+φ2+φ1)](15)\mathcal{H}^{\prime\prime}_{c}=J_{1}\overline{S}_{c}^{2}\cos(h_{1}q_{\mathrm{cyc}}/2)\sum_{n}\left[\cos\left((h_{2}-h_{1})q_{\mathrm{cyc}}(n+1/2)+\varphi_{2}-\varphi_{1}\right)+\cos\left((h_{2}+h_{1})q_{\mathrm{cyc}}(n+1/2)+\varphi_{2}+\varphi_{1}\right)\right] \tag*{(15)}

and the two cases p=h1⋅h2=+1p=h_{1}\cdot h_{2}=+1 and p=−1p=-1 yield

Ep=+1′′=J1N(S‾c2+S‾b2)cos⁡(qcyc/2)cos⁡(φ2−φ1)(16)E^{\prime\prime}_{p=+1}=J_{1}N\left(\overline{S}_{c}^{2}+\overline{S}_{b}^{2}\right)\cos(q_{\mathrm{cyc}}/2)\cos(\varphi_{2}-\varphi_{1}) \tag*{(16)}
Ep=−1′′=J1N(S‾c2−S‾b2)cos⁡(qcyc/2)cos⁡(φ2+φ1)(17)E^{\prime\prime}_{p=-1}=J_{1}N\left(\overline{S}_{c}^{2}-\overline{S}_{b}^{2}\right)\cos(q_{\mathrm{cyc}}/2)\cos(\varphi_{2}+\varphi_{1}) \tag*{(17)}

These terms in Eq. (16), (17) are independent of znz_{n}; especially for p=+1p=+1, optimizing δ=φ2−φ1\delta=\varphi_{2}-\varphi_{1} yields a favorable energy contribution for any given J1J_{1}, ECDWE_{\mathrm{CDW}}.

Expressions for scattering intensities

We review the expressions used to calculate neutron scattering intensities from atomic and magnetic structures of DyTe3_3, which serve to define a variety of parameters (such as the phase shift δ\delta) used in the discussion of the main text. Working with the triple-axis diffractometer PONTA-5G at JRR-3 research reactor, we fit ω\omega or θ−2θ\theta-2\theta scans of neutron intensities with Gaussian profiles and calculate the total observed intensity IobsI_{\mathrm{obs}} for each reflection, taking the Q\mathbf{Q}-dependence of the peak shape into account (Figure E10). From this, we calculate the observed structure factor as

Fobs(Q)=Iobs(Q)L(2θ)(18)F_{\mathrm{obs}}(\mathbf{Q})=\sqrt{\frac{I_{\mathrm{obs}}(\mathbf{Q})}{\mathcal{L}(2\theta)}} \tag*{(18)}

where L(2θ)=λ3/sin⁡(2θ)\mathcal{L}(2\theta)=\lambda^{3}/\sin(2\theta) is the Lorentz factor, λ\lambda is the wavelength of the monochromatized neutrons, and 2θ2\theta is the scattering angle.

To reproduce this Fobs(Q)F_{\mathrm{obs}}(\mathbf{Q}) quantitatively, we start from the expression for the differential cross-sections for nuclear and magnetic scattering, i.e. the beam intensity scattered into a solid angle dΩ\mathrm{d}\Omega corresponding to the direction of the momentum transfer Q\mathbf{Q} of magnitude Q=∣Q∣Q = |\mathbf{Q}| and direction Q^=Q/Q\hat{\mathbf{Q}} = \mathbf{Q}/Q

(dσdΩ)calN≡IcalN=Φ∣∑j∈latticebjexp⁡(iQ⋅rj)∣2≡Φ∣FcalN(Q)∣2∑1,l′exp⁡[iQ⋅(l−l′)](19)\left(\frac{\mathrm{d}\sigma}{\mathrm{d}\Omega}\right)_{\mathrm{cal}}^{N} \equiv I_{\mathrm{cal}}^{N} = \Phi\left|\sum_{j\in\mathrm{lattice}} b_j \exp\left(\mathrm{i}\mathbf{Q}\cdot\mathbf{r}_j\right)\right|^2 \equiv\Phi\left|F_{\mathrm{cal}}^{N}(\mathbf{Q})\right|^2 \sum_{1,l'} \exp\left[\mathrm{i}\mathbf{Q}\cdot(\mathbf{l}-\mathbf{l}')\right] \tag*{(19)}
(dσdΩ)calM≡IcalM=Φ∣−2.7∑j∈latticefmag(Q)m⊥,jexp⁡(iQ⋅rj)∣2(20)\left(\frac{\mathrm{d}\sigma}{\mathrm{d}\Omega}\right)_{\mathrm{cal}}^{M} \equiv I_{\mathrm{cal}}^{M} = \Phi\left|-2.7\sum_{j\in\mathrm{lattice}} f_{\mathrm{mag}}(\mathbf{Q})\mathbf{m}_{\perp,j}\exp\left(\mathrm{i}\mathbf{Q}\cdot\mathbf{r}_j\right)\right|^2 \tag*{(20)}

Here, we have introduced the neutron flux Φ\Phi, the position of each atom on the lattice rj\mathbf{r}_j, the nuclear scattering length bjb_j, the unit cell coordinates l,l′\mathbf{l},\mathbf{l}', the magnetic form factor fmag(Q)f_{\mathrm{mag}}(\mathbf{Q}) evaluated from an analytic expression [72], and the component of the magnetic moment at site jj that is perpendicular to the momentum transfer

m⊥,j(Q)=mj−(mj⋅Q^)⋅Q^(21)\mathbf{m}_{\perp,j}(\mathbf{Q}) = \mathbf{m}_j - \left(\mathbf{m}_j\cdot\hat{\mathbf{Q}}\right)\cdot\hat{\mathbf{Q}} \tag*{(21)}

Note the scattering lengths for Dy3+\mathrm{Dy}^{3+} (16.9 fm) and Te (5.8 fm), and that only magnetic ions contribute to Eq. (20).

For nuclear scattering, we use [73]

FcalN(Q)=∑d∈c.u.c.bjexp⁡(iQ⋅d)(22)F_{\mathrm{cal}}^{N}(\mathbf{Q}) = \sum_{\mathbf{d}\in\mathrm{c.u.c.}} b_j\exp\left(\mathrm{i}\mathbf{Q}\cdot\mathbf{d}\right) \tag*{(22)}
Φ∑l,l′exp⁡[iQ⋅(l−l′)]=ΦN(2π)3ν0δ(3)(Q−G)(23)\Phi\sum_{\mathbf{l},\mathbf{l}'}\exp\left[\mathrm{i}\mathbf{Q}\cdot(\mathbf{l}-\mathbf{l}')\right] = \Phi N\frac{(2\pi)^3}{\nu_0}\delta^{(3)}(\mathbf{Q}-\mathbf{G}) \tag*{(23)}

for a reciprocal lattice vector G\mathbf{G} and using rj=d+l\mathbf{r}_j=\mathbf{d}+\mathbf{l}, where the former is a coordinate within the crystallographic unit cell, and the latter labels the origin of each unit cell. Φ\Phi, NN, FcalNF_{\mathrm{cal}}^{N}, are the flux of incident neutrons, the number of crystallographic unit cells (c.u.c., of volume ν0\nu_0) in the sample, and the nuclear structure factor. We find good agreement of the experimental scattering data and model when using the atomic positions from the high-temperature CmcmCmcm space group of DyTe3_3 [74]. In reality, the formation of charge order below TCDW≈320T_{\mathrm{CDW}}\approx320 K lowers the symmetry, as discussed by Malliakas et al. in Refs. [18, 75].

Then, a scale factor ss is defined by equating to the experimentally observed intensity,

∣FobsN(Q)∣2=ΦN(2π)3ν0∣FcalN(Q)∣2=s∣FcalN(Q)∣2(24)\left|F_{\mathrm{obs}}^{N}(\mathbf{Q})\right|^2 = \Phi N\frac{(2\pi)^3}{\nu_0}\left|F_{\mathrm{cal}}^{N}(\mathbf{Q})\right|^2 = s\left|F_{\mathrm{cal}}^{N}(\mathbf{Q})\right|^2 \tag*{(24)}

as shown in Figure E7.

Image not extracted

Figure E7. Nuclear reflections observed in neutron scattering, consistent with orthorhombic crystal structure of DyTe3. The observed structure factor FobsF_{\mathrm{obs}} is compared to model calculations, FcalF_{\mathrm{cal}}, in space group CmcmCmcm as detailed in section E2. We use the reliability index for the structure factor, R=∑∣Fobs−Fcal∣/∑∣Fobs∣R = \sum\lvert F_{\mathrm{obs}} - F_{\mathrm{cal}}\rvert/\sum\lvert F_{\mathrm{obs}}\rvert, where the sums are over all data points; R=0.0328R = 0.0328 for sample B (a) and R=0.074R = 0.074 for sample A (b) indicate good agreement of model and fit. Note that ω\omega-scans without absorption correction are used to obtain the integrated intensities for sample A.

Next, for magnetic scattering from a structure with lattice-commensurate magnetic order, we use Eq. (23) with a larger unit cell (reduced set of G\mathbf{G}) and equate

∣FobsM(Q)∣2=IobsM(Q)/L(2θ)=IcalM(Q)=ΦNM(2π)3νM∣FcalM(Q)∣2≡s∣FcalMx∣2(25)\left|F_{\mathrm{obs}}^{M}(\mathbf{Q})\right|^2 = I_{\mathrm{obs}}^{M}(\mathbf{Q})/\mathcal{L}(2\theta) = I_{\mathrm{cal}}^{M}(\mathbf{Q}) = \Phi N_{M}\frac{(2\pi)^3}{\nu_{M}}\left|F_{\mathrm{cal}}^{M}(\mathbf{Q})\right|^2 \equiv s\left|\frac{F_{\mathrm{cal}}^{M}}{x}\right|^2 \tag*{(25)}
FcalM(Q)=−2.7fmag(Q)∑j∈m.u.c.m1,jexp⁡(iQ⋅rj)(26)\mathbf{F}_{\mathrm{cal}}^{M}(\mathbf{Q}) = -2.7f_{\mathrm{mag}}(Q)\sum_{j\in\mathrm{m.u.c.}}\mathbf{m}_{1,j}\exp(i\mathbf{Q}\cdot\mathbf{r}_{j}) \tag*{(26)}

We introduced the calculated magnetic structure factor FcalM\mathbf{F}_{\mathrm{cal}}^{M}, which has three complex components Fcal,xMF_{\mathrm{cal},x}^{M}, Fcal,yMF_{\mathrm{cal},y}^{M}, Fcal,zMF_{\mathrm{cal},z}^{M}; the volume of the magnetic unit cell (m.u.c.) νM=x⋅ν0\nu_{M}=x\cdot\nu_{0}, the magnetic form factor of Dy3+^{3+}, fmagf_{\mathrm{mag}}, and the number NM=N/xN_{M}=N/x of m.u.c. in the sample. The prefactor (−2.7-2.7) describes the scattering length of the electron, and the sum is now over all magnetic (dysprosium) ions in the m.u.c.

In case of two domains of this commensurate order, as relevant for the analysis in DyTe3_3, we have

IcalM=ΦN2(2π)3ν0∣FcalM1x∣2+ΦN2(2π)3ν0∣FcalM2x∣2=s2(∣FcalM1x∣2+∣FcalM2x∣2)(27)I_{\mathrm{cal}}^{M}=\Phi\frac{N}{2}\frac{(2\pi)^3}{\nu_{0}}\left|\frac{F_{\mathrm{cal}}^{M1}}{x}\right|^2+\Phi\frac{N}{2}\frac{(2\pi)^3}{\nu_{0}}\left|\frac{F_{\mathrm{cal}}^{M2}}{x}\right|^2=\frac{s}{2}\left(\left|\frac{F_{\mathrm{cal}}^{M1}}{x}\right|^2+\left|\frac{F_{\mathrm{cal}}^{M2}}{x}\right|^2\right) \tag*{(27)}

where FcalMk\mathbf{F}_{\mathrm{cal}}^{Mk}, k=1,2k=1,2 are the magnetic structure factors for two domains.

Symmetry and structure factor: commensurate component in phase I

Symmetry consideration (commensurate)

Section E1 discusses oscillatory terms that are allowed in the Hamiltonian due to local symmetry breaking from the charge-density wave (CDW). In this section, we focus on global symmetries of DyTe3_3 and their lowering by magnetic order. We start from space group CmmmCmmm, ignoring the incommensurate CDW at first, and briefly consider the effect of the CDW on global symmetries at the end of the section.

In polarized neutron scattering (PNS), orthorhombic symmetry of the crystal structure allows us to set the scattering plane to HK0HK0, with separation of three orthogonal magnetization components mam_a, mbm_b, and mcm_c. The PNS data strictly constrains the commensurate moment to be along the aa-axis, and the observation of reflections of the type (0, odd, 0.5) – while (0, even, 0.5) are absent –, establishes a unit cell with eight magnetic dysprosium ions, vanishing net magnetization along the aa-axis, as well as opposite directions for moments separated by a distance b/2b/2 along the bb-axis (Fig. E3c). Four free parameters remain, namely

Image not extracted

Figure E2. Space group CmcmCmcm and its symmetries in the (000)+(000)+ set, as well as breaking of bond-symmetry by AFM commensurate order. Only dysprosium ions are shown (blue), including two ions outside the crystallographic unit cell (right) to illustrate the structural zigzag chain feature. a, Mirror plane Ma\mathcal{M}_{a}. b, Mirror plane Mc\mathcal{M}_{c}. c, Glide mirror plane Gb\mathcal{G}_{b} with c/2c/2 translation along the cc-axis (black arrow). d, Two-fold rotation C2aC_{2a} through the inversion center I=(12,12,12)\mathcal{I} = \left(\frac{1}{2},\frac{1}{2},\frac{1}{2}\right). e, Two-fold rotation C2bC_{2b} through a position shifted by (0,0,14)\left(0,0,\frac{1}{4}\right) from I\mathcal{I}. f, Screw rotation C2c×(0,0,12)C_{2c} \times\left(0,0,\frac{1}{2}\right) with c/2c/2 shift along the cc-axis. g Base centering operation CC, i.e., a translation by (12,12,0)\left(\frac{1}{2},\frac{1}{2},0\right). h In the presence of commensurate order qAFM=(0,b∗,0.5c∗)\mathbf{q}_{\mathrm{AFM}} = (0,b^{*},0.5c^{*}), the magnetic space group symmetry is lowered to Cc2/mCc2/m: Mc\mathcal{M}_{c} and C2bC_{2b} are broken (grey), while Ma\mathcal{M}_{a} survives (red). The equivalence of bonds on the effective zigzag chain (main text) is lifted in presence of uudduudd or udduuddu order (black double and single bonds).

the lengths of four magnetic moments mam_a in a DyTe slab spanning two crystallographic unit cells, i.e. the upper zigzag chain in Figure E3.

Unit cell and magnetic domains for commensurate spin component in phase I of DyTe3

Figure E3. Unit cell, magnetic domains for commensurate spin component, and illustration for structure factor calculation in phase I of DyTe3DyTe_3 (antiferromagnetic part, AFM). a, b, Primitive magnetic unit cell (red boundary, monoclinic) in two AFM domains of DyTe3DyTe_3, with base centering translation vector T1\mathcal{T}_1 and CcC_c translation vector (plus time reversal) T2\mathcal{T}_2. The base centering operation of the parent structure CmcmCmcm (C), and its inversion center (I), are also indicated by a violet arrow and cross, respectively. Only magnetic sites are shown, and sites with different xx-positions have been projected onto a single plane. See section E3. c, Conventional unit cell for two AFM domains in phase I. Circles with dot (with cross) signify magnetic moment pointing into (out of) the plane of the figure. The bb-axis spacing of magnetic sites in a bilayer (grey shading) is approximately b/6b/6. Inversion centers I\mathcal{I} are indicated by violet crosses. d, Two sets of sites (black, white circles) of magnetic ions in the unit cell. In the calculation of the structure factor for the combined scattering of two equally populated domains A, B according to Eq. (28), the only contributions are from terms ∼exp⁡(iQ⋅(ri−rj))\sim\exp(\mathrm{i}\mathbf{Q}\cdot(\mathbf{r}_i-\mathbf{r}_j)) for sites i,ji,j in the same set.

The full Hermann–Mauguin symbol for orthorhombic space group CmcmCmcm (SG63) is C2/m2/c21/mC 2/m 2/c 2_1/m, and qAFM=(0,1,0.5)\mathbf{q}_{\mathrm{AFM}} = (0,1,0.5) does not break any of these symmetries. Using the k-Subgroupsmag tool of the Bilbao crystallographic server [76], we determined the highest-symmetry magnetic subgroups of CmcmCmcm that are consistent with qAFM\mathbf{q}_{\mathrm{AFM}}. In (black-and-white-type) Belov-Neronova-Smirnova notation: Iama2I_a ma2, Cc2/cCc2/c, and Cc2/mCc2/m. From the transformation matrices provided by this tool, we also find the lattice vectors a′\mathbf{a}', b′\mathbf{b}', c′\mathbf{c}' in terms of the lattice vectors a\mathbf{a}, b\mathbf{b}, c\mathbf{c} of the ’parent’ CmcmCmcm: (−2c,a,−b)(-2\mathbf{c},\mathbf{a},-\mathbf{b}), (−b−2c,a,2c)(-\mathbf{b}-2\mathbf{c},\mathbf{a},2\mathbf{c}), and (b+2c,−a,2c)(\mathbf{b}+2\mathbf{c},-\mathbf{a},2\mathbf{c}), for the three abovementioned symbols. Here, for example, b=eb⋅b\mathbf{b}=\mathbf{e}_b\cdot b for basis (unit) vector eb\mathbf{e}_b and lattice constant bb (and so on).

We rule out Iama2I_a ma2 and Cc2/cCc2/c for the commensurate order in DyTe3_3. The former has a Ma′\mathcal{M}_{a'} mirror plane, which is perpendicular to ec\mathbf{e}_c of the CmcmCmcm cell and located on the Dy-sites. A moment mam_a on the Dy-site is inconsistent with this mirror plane. The latter space group has a glide mirror plane Gb′\mathcal{G}'_b, consisting of a mirror operation perpendicular to ea\mathbf{e}_a combined with a lattice translation by c\mathbf{c}. The moment direction mam_a at a given site is unchanged under Gb′\mathcal{G}'_b, but translated by half the length of the magnetic unit cell. Given there are merely four sites in the upper DyTe slab of the magnetic unit cell, this operation requires orders of the type udud, and is thus inconsistent with the expansion of the unit cell along ec\mathbf{e}_c that is implied by qAFM\mathbf{q}_{\mathrm{AFM}}.

We focus on monoclinic, centrosymmetric Cc2/mCc2/m, where CcC_c includes base centering as a translation T1=(1/2,1/2,0)\mathcal{T}_1=(1/2,1/2,0) and T2′\mathcal{T}'_2, i.e. the translation T2=(0,0,1/2)\mathcal{T}_2=(0,0,1/2) combined with time reversal. There are two domains A and B with basis vectors (b+2c,−a,2c)(\mathbf{b}+2\mathbf{c},-\mathbf{a},2\mathbf{c}) and (−b+2c,a,2c)(-\mathbf{b}+2\mathbf{c},\mathbf{a},2\mathbf{c}); i.e., they are characterized by a reversal of the monoclinic tilt (Figure E3). Only the mirror plane perpendicular to ea\mathbf{e}_a remains intact, the number of inversion centers is halfed as compared to CmcmCmcm, and the broken cc-mirror relates the two possible domains A and B depicted in Figure E3c.

Note: A previous x-ray scattering study reports the superspace group C2cm(00γ)000C2cm(00\gamma)000 for the CDW state in DyTe3_3 [75]. In average space group C2cmC2cm (No. 40), the Ma\mathcal{M}_a mirror of CmcmCmcm is already broken. Starting from this lower-symmetry symbol, analogous discussion leads to Cc2Cc2 for the commensurate component of the magnetic order in phase I.

We consider the possibility of further symmetry lowering, as caused by non-uniform mag-

netic moment length mam_a. Magnetic space group Cc2/mC_c2/m has T2\mathcal{T}_2, which ensures that moments mam_a in the same layer α\alpha point in opposite directions, but have the same length. Instead of uudduudd (up-up-down-down), this leaves (ushortulongdshortdlong)(u_{\mathrm{short}}u_{\mathrm{long}}d_{\mathrm{short}}d_{\mathrm{long}}) as a viable configuration, but C2aC_{2a} rotation symmetry interrelates layers α=1,4\alpha=1,4 and eliminates this possibility. The neutron data are well described by Cc2/mC_c2/m (or Cc2C_c2, which leads to the same commensurate structure models), so that a further symmetry lowering – relaxing the uniformity constraint on mam_a – is deemed unnecessary.

Structure factor calculation (commensurate)

We now derive explicit expressions for the scattered intensity at commensurate (AFM) reflections in momentum space, in phase I of DyTe3_3.

For momentum transfer Q\mathbf{Q} in the (0KL)(0KL) plane, we define a normalized moment for each domain m⊥,j=mafA,B(rj)ea\mathbf{m}_{\perp,j}=m_a f_{A,B}(\mathbf{r}_j)\mathbf{e}_a, where fA,B=±1f_{A,B}=\pm1. Assuming equal population of domains A and B, Eq. (27) gives

IcalM=s8(−2.7fmag(Q)ma)2∑i,j=1,…,8exp⁡[iQ⋅(ri−rj)]⋅[fA(ri)fA∗(rj)+fB(ri)fB∗(rj)](28)I_{\mathrm{cal}}^{M}=\frac{s}{8}\left(-2.7f_{\mathrm{mag}}(\mathbf{Q})m_a\right)^2 \sum_{i,j=1,\ldots,8}\exp\left[\mathrm{i}\mathbf{Q}\cdot(\mathbf{r}_i-\mathbf{r}_j)\right]\cdot\left[f_A(\mathbf{r}_i)f_A^*(\mathbf{r}_j)+f_B(\mathbf{r}_i)f_B^*(\mathbf{r}_j)\right] \tag*{(28)}

where the magnetic sites i,ji,j are labeled in Figure E3 d (not Figure 2). In the latter edgy brackets, the two terms cancel (sum to ±2\pm2) when the relationship between magnetic moments at site i,ji,j is opposite (the same) in domains A, B. In addition to four pairs of type i=ji=j, examination of Figure E3 d shows that, for the purpose of evaluating Eq. (28), the magnetic sites can be split into two sets 1 and 2: the non-vanishing terms in the sum correspond to pairs of magnetic moments that are just above / below one another, or two sites apart along the cc-direction.

The relative distances between sites are the same for set 1 and 2, and there are three types of site pairs i,ji,j in each set:

ri−rj=±cec(29)\mathbf{r}_i-\mathbf{r}_j=\pm c\mathbf{e}_c \tag*{(29)}
ri−rj=±(b/2)eb(30)\mathbf{r}_i-\mathbf{r}_j=\pm(b/2)\mathbf{e}_b \tag*{(30)}
ri−rj=±(cec−(b/2)eb)(31)\mathbf{r}_i-\mathbf{r}_j=\pm\left(c\mathbf{e}_c-(b/2)\mathbf{e}_b\right) \tag*{(31)}

with both ±\pm appearing in the sum of Eq. (28). The moments are antiparallel for all pairs of sites, [fA(ri)fA∗(rj)+fB(ri)fB∗(rj)]=−2[f_A(\mathbf{r}_i)f_A^*(\mathbf{r}_j)+f_B(\mathbf{r}_i)f_B^*(\mathbf{r}_j)]=-2, except the pairs in Eq. (31). We specialize to the (0,Kb∗,Lc∗+qAFM)(0,Kb^*,Lc^*+q_{\mathrm{AFM}}) scattering plane, so that

Q=K2πbeb+(L+0.5)2πcec(32)\mathbf{Q}=K\frac{2\pi}{b}\mathbf{e}_b+(L+0.5)\frac{2\pi}{c}\mathbf{e}_c \tag*{(32)}

as in our experiments, and obtain

∑i≠j,set1exp⁡(iQ⋅(ri−rj))=4cos⁡(Qcc)+4cos⁡(Qbb2)−4cos⁡(Qcc)cos⁡(Qbb2)=−4+4⋅(−1)K+4⋅(−1)K=4(2(−1)K−1)(33)\sum_{i\ne j,\mathrm{set1}}\exp\left(\mathrm{i}\mathbf{Q}\cdot(\mathbf{r}_i-\mathbf{r}_j)\right)=4\cos(Q_c c)+4\cos\left(Q_b\frac{b}{2}\right)-4\cos(Q_c c)\cos\left(Q_b\frac{b}{2}\right) = -4+4\cdot(-1)^K+4\cdot(-1)^K=4\left(2(-1)^K-1\right) \tag*{(33)}

which takes the values −12-12 (+4+4) for K=oddK=\mathrm{odd} (even\mathrm{even}), independent of LL. Then, ∣FobsM∣≡0\lvert F_{\mathrm{obs}}^M\rvert\equiv0 when KK is even, and for K=oddK=\mathrm{odd},

IcalM=s8(−2.7fmag(Q)ma)2(2⋅8−2⋅2⋅(−12))=8s(−2.7fmag(Q)ma)2(34)I_{\mathrm{cal}}^M=\frac{s}{8}\left(-2.7f_{\mathrm{mag}}(Q)m_a\right)^2\left(2\cdot8-2\cdot2\cdot(-12)\right) =8s\left(-2.7f_{\mathrm{mag}}(Q)m_a\right)^2 \tag*{(34)}

without any explicit dependence on KK and LL both (but note the magnetic form factor). The above discussion shows that contributions from domains A, B partially cancel each other, meaning that the ratio of domains can be refined from the data.

Symmetry and structure factor: incommensurate component in phase I

Symmetry consideration (incommensurate)

Starting from Eqs. (12), (13), and given antiferromagnetic coupling between structural bilayers in DyTe3_3, as enforced by K=oddK=\mathrm{odd} in (0,K,Lcyc)(0,K,L_{\mathrm{cyc}}), the phase relation between layers α=1,3\alpha=1,3 and 2,42,4 is fixed to π\pi and h1=h3h_1=h_3, h2=h4h_2=h_4. The remaining free parameters, besides Sˉb\bar{S}_b, Sˉc\bar{S}_c, are a single phase δ=φ1−φ2=φ3−φ4\delta=\varphi_1-\varphi_2=\varphi_3-\varphi_4 and the parameter p=h1⋅h2=±1p=h_1\cdot h_2=\pm1, which characterizes the relative helicity in two adjacent layers. We set a single (possibly distorted), incommensurate cycloid into the uppermost layer α=1\alpha=1 with φ1=0\varphi_1=0, and – enforced by K=oddK=\mathrm{odd} – a copy with φ3=π\varphi_3=\pi into the third layer. We then approximate the incommensurate order by an (arbitrarily large) supercell, and discuss the symmetries that leave the magnetic moments unchanged. From this, we deduce constraints on pp and δ\delta.

Starting from Cc2/mCc2/m, the cycloid in α=1\alpha= 1 removes the Ma\mathcal{M}_a mirror operation. Figure E4a shows the point group symmetries of a cycloid, with one of the m′m' operations perpendicular to ea\mathbf{e}_a in DyTe3_3; but the (mirror ×\times time reversal) operation is not consistent with the commensurate component of phase I. Instead, a cc-glide appears in the place of Ma\mathcal{M}_a (in the frame of CmcmCmcm). The translations T1\mathcal{T}_1 and T2′\mathcal{T}'_2 of CcC_c (Figure E3) are broken by the cycloids in α=1,3\alpha=1,3. Without further symmetry lowering, the average magnetic space group is C2/cC2/c, where – somewhere along the cc-direction – a global inversion center survives. Due to the remaining C2aC_{2a} symmetry, the four layers are stacked as ha=(1,−1,1,−1)\mathbf{h}_a=(1,-1,1,-1), where h=(h1,h2,h3,h4)\mathbf{h}=(h_1,h_2,h_3,h_4); recall that h1=h3h_1=h_3 and h2=h4h_2=h_4 (Figure E4b). Importantly, the configuration hb=(−1,1,−1,1)\mathbf{h}_b=(-1,1,-1,1) is not related to ha\mathbf{h}_a by a broken symmetry and does not have the same energy as ha\mathbf{h}_a; the energy difference between these two is largely determined by the local Dzyaloshinskii-Moriya (local DMI) interaction on the J2J_2 bond of the zigzag chain (Figure 1), i.e. the local DMI between nearest neighbors on a single square net of the DyTe slab. The competition between inter-layer Heisenberg exchange J1J_1, Eq. (10), and intra-layer local DMI determines the choice of co-propagating or counter-propagating cycloids, and hence the average magnetic space group.

Figure E4. Symmetry considerations regarding incommensurate magnetic order in phase I, for DyTe3. a, Magnetic point group symmetries of a cycloid. b, A magnetic texture with P2P2 symmetry and ordering vector qcyc\mathbf{q}_{\mathrm{cyc}}. Red arrows: Setting a cycloid in the uppermost layer α=1\alpha= 1 with helicity h1=+1h_1 = +1, the α=3\alpha= 3 layer’s structure with h3=+1h_3 = +1 is imposed by antiferromagnetic coupling (K=1K = 1). Violet arrows: Applying C2aC_{2a} symmetry enforces a counter-propagating cycloid in layers α=2,4\alpha= 2, 4, with h2=h4=−1h_2 = h_4 = -1, where the phase δ=φ2−φ1\delta= \varphi_2 - \varphi_1 is fixed to 38∘38^\circ. The texture in layer α=2\alpha= 2 can be inferred from AFM coupling to α=4\alpha= 4. c, Switching between helicity domains in the CcCc model with uniform helicity h=±(1,1,1,1)\mathbf{h} = \pm(1,1,1,1); for simplicity of the illustration, the phase difference is set to δ=38∘\delta= 38^\circ. Grey arrows: assuming broken C2aC_{2a} symmetry, uniform helicity and adjustable phase shift δ\delta are allowed. The C2aC_{2a} operation (violet arrow) maps h↔−h\mathbf{h} \leftrightarrow-\mathbf{h}, while leaving δ\delta unchanged. (Nearly) collinear blocks, (purple highlight) remain coupled. d, Simultaneous switching of δ\delta and h\mathbf{h} via glide mirror Gb\mathcal{G}_b, and switching of δ\delta without h\mathbf{h} by mirror Mc\mathcal{M}_c. Highlighted ovals mark pairs of spins with nearly collinear alignment in the bcbc plane. Again, δ=±38∘\delta= \pm38^\circ is used for the purpose of this illustration, although the experimental value is 79∘79^\circ.

Image not extracted

Figure E5. Two ’δ\delta-domains’ (A′ and B′), consistent with the qAFM=(0,b∗,0.5c∗)q_{\mathrm{AFM}}=(0,b^*,0.5c^*) ordering vector, for the incommensurate part of the magnetic order in phase I of DyTe3_3. The value of the phase shift δ=79∘\delta=79^\circ between sheets in a bilayer is adjusted to minimize the reliability index RR for the structure factor FF. Grey arrows indicate the original moment direction that is expected without any phase shift δ\delta. Note that there is a phase shift of π\pi between upper and lower Dy-bilayer, and that reversing the sign of the helicity vector h=(h1,h2,h3,h4)\mathbf{h}=(h_1,h_2,h_3,h_4), as defined in section E4, does not affect the value of scattering intensities; η\eta is the angle between the bcbc-component of the magnetic moment and the crystallographic bb-axis.

For the ha=(1,−1,1,−1)\mathbf{h}_a=(1,-1,1,-1) stack, a degree of freedom remains regarding the phase relation of cycloids in two layers of a DyTe slab. In particular, the phase shift angle δ\delta determines whether blocks of parallel moments point along eb\mathbf{e}_b and blocks of antiparallel moments point along ec\mathbf{e}_c, or vice versa. Absent further symmetry lowering, the remaining C2aC_{2a} symmetry in C2/cC2/c fixes h4=−h1h_4=-h_1, with δ0=±qcyc⋅0.5c≈±0.207π=±37.8∘\delta_0=\pm q_{\mathrm{cyc}}\cdot0.5c\approx\pm0.207\pi=\pm37.8^{\circ}, where the sign depends on the position of the inversion center in Cc2/mCc2/m, i.e., on domain A or B of the AFM order. We have derived an analytic expression for the magnetic structure factor of counter-propagating cycloids in DyTe3_3, section E4 D, and consider the alternative models of high-symmetry, counter-propagating cycloids in that section.

For h=(1,−1,1,−1)\mathbf{h}=(1,-1,1,-1), we can further lower symmetry to CcCc (on average, space group number 9) by shifting δ\delta away from δ0\delta_0, thus breaking C2aC_{2a}. However, the spontaneous formation of helimagnetism with a unique sense of rotation is common in zigzag chain magnets, and more generally in systems where two or more sublattices are connected by space inversion, a twofold screw axis along the chain, and / or a glide mirror. Examples are Ni3_3V2_2O8_8 [77], CuO [78], and the zigzag chain magnet MnWO4_4 [79], which all host single-sense helices under these conditions. Thus, we consider the cycloid of uniform helicity h(1)=(1,1,1,1)\mathbf{h}^{(1)}=(1,1,1,1) (average space group CcCc), which allows the spin system to adjust δ\delta to minimize inter-chain exchange energy (section E1). We are left with two helicity-domains, h(1)\mathbf{h}^{(1)} and h(2)=−h(1)\mathbf{h}^{(2)}=-\mathbf{h}^{(1)} for each AFM domain (uudduudd or udduuddu). These are related by C2aC_{2a}, which reverses the rotation sense of the cycloids, but maintains the same AFM domain and the same phase relation δ\delta for a given bond (Figure E4c). As the symmetry is reduced to CcCc in each domain, there is no constraint on the number value of δ\delta. The discussion remains qualitatively unchanged if the commensurate part has magnetic space group Cc2Cc2.

Figure E2h demonstrates the inequivalence of bonds on the zigzag chain, caused by the commensurate magnetic structure component, which is essential for symmetry reduction and for allowing off-diagonal anisotropy terms such as SnaSncS_n^{a}S_n^{c} in the Hamiltonian of section E1. In particular, δ\delta becomes a refinable parameter only when the average symmetry is lowered to CcCc.

Structure factor calculation (incommensurate)

We translate the the spin-chain model, Eqs. (12), (13), into expressions suitable for structure factor calculations. Instead of the full, three-dimensional propagation vector qcyc=(0,b∗,0.207c∗)\mathbf{q}_{\mathrm{cyc}}=(0,b^*,0.207c^*), we use qcyc′=(0,0,0.207c∗)\mathbf{q}'_{\mathrm{cyc}}=(0,0,0.207c^*), which helps to define the equations and relative phases of magnetic layers in language that is consistent with the main text and with section E4 A. The magnetic sites are labeled by index jj as in Figure 2, rj=l+d\mathbf{r}_j=\mathbf{l}+\mathbf{d}, where l\mathbf{l} and d\mathbf{d} label the origin of the unit cell and the intra-cell coordinate of the site, respectively. As there is only one magnetic (dysprosium) site per layer α\alpha [defined in Eqs. (12), (13)], we replace the layer index α\alpha by dd and write

mj=2Xcos⁡[hd⋅qcyc′⋅(d+l)+φd]−2Ysin⁡[hd⋅qcyc′⋅(d+l)+φd]=(X+ıY)exp⁡(ıhd⋅qcyc′⋅(d+l)+ıφd)++(X−ıY)exp⁡(−ıhd⋅qcyc′⋅(d+l)−ıφd)(35)\begin{aligned} \mathbf{m}_j=2\mathbf{X}\cos\left[h_d\cdot\mathbf{q}'_{\mathrm{cyc}}\cdot(\mathbf{d}+\mathbf{l})+\varphi_d\right]-2\mathbf{Y}\sin\left[h_d\cdot\mathbf{q}'_{\mathrm{cyc}}\cdot(\mathbf{d}+\mathbf{l})+\varphi_d\right] \\ &=(\mathbf{X}+\imath\mathbf{Y})\exp\left(\imath h_d\cdot\mathbf{q}'_{\mathrm{cyc}}\cdot(\mathbf{d}+\mathbf{l})+\imath\varphi_d\right)+ \\ &+(\mathbf{X}-\imath\mathbf{Y})\exp\left(-\imath h_d\cdot\mathbf{q}'_{\mathrm{cyc}}\cdot(\mathbf{d}+\mathbf{l})-\imath\varphi_d\right) \tag*{(35)} \end{aligned}

where X=XeX\mathbf{X}=X\mathbf{e}_X, Y=YeY\mathbf{Y}=Y\mathbf{e}_Y with two orthogonal unit vectors eX\mathbf{e}_X, eY\mathbf{e}_Y. As before, the phase degree of freedom φd\varphi_d is specific to each layer index d(rj)=1,2,3,4d(\mathbf{r}_j)=1,2,3,4, and the prefactors hd=±1h_d=\pm1 indicate the helicity (sense of rotation, or handedness) in a layer. Based on the discussion in section E4 A, we use φ1=0\varphi_1=0, φ2=±δ\varphi_2=\pm\delta, φ3=π\varphi_3=\pi, and φ4=π±δ\varphi_4=\pi\pm\delta, where ±\pm indicates two magnetic domains of the phase shift angle δ\delta.

We insert this into Eq. (20), drop the helicity factor hdh_d, and define X⊥=X−X⋅Q^\mathbf{X}_{\perp}=\mathbf{X}-\mathbf{X}\cdot\hat{\mathbf{Q}} with a normal vector Q^=Q/∣Q∣\hat{\mathbf{Q}}=\mathbf{Q}/|\mathbf{Q}|. The quantity Y⊥\mathbf{Y}_{\perp} is defined analogously, and we are careful to consider the structure factor in the incommensurate case as a sum over the full lattice, not over a single magnetic unit cell,

IcalM=Φ∣−2.7fmag(Q)∣2(X⊥2+Y⊥2)⋅[∣F~+(Q)∣2∑l,l′exp⁡(ı(Q+qcyc′)⋅(l−l′))+∣F~−(Q)∣2∑l,l′exp⁡(ı(Q−qcyc′)⋅(l−l′))](36)\begin{aligned} I_{\mathrm{cal}}^M=\Phi\left|-2.7f_{\mathrm{mag}}(\mathbf{Q})\right|^2\left(\mathbf{X}_{\perp}^2+\mathbf{Y}_{\perp}^2\right) \\ &\cdot\left[\left|\widetilde{F}^{+}(\mathbf{Q})\right|^2\sum_{\mathbf{l},\mathbf{l}'}\exp\left(\imath(\mathbf{Q}+\mathbf{q}'_{\mathrm{cyc}})\cdot(\mathbf{l}-\mathbf{l}')\right)+\left|\widetilde{F}^{-}(\mathbf{Q})\right|^2\sum_{\mathbf{l},\mathbf{l}'}\exp\left(\imath(\mathbf{Q}-\mathbf{q}'_{\mathrm{cyc}})\cdot(\mathbf{l}-\mathbf{l}')\right)\right] \tag*{(36)} \end{aligned}
F~+(Q)=∑d∈c.u.c.exp⁡(ı(Q+qcyc′)⋅d+ıφd)(37)\widetilde{F}^{+}(\mathbf{Q})=\sum_{\mathbf{d}\in\mathrm{c.u.c.}}\exp\left(\imath(\mathbf{Q}+\mathbf{q}'_{\mathrm{cyc}})\cdot\mathbf{d}+\imath\varphi_d\right) \tag*{(37)}
F~−(Q)=∑d∈c.u.c.exp⁡(ı(Q−qcyc′)⋅d−ıφd)(38)\widetilde{F}^{-}(\mathbf{Q})=\sum_{\mathbf{d}\in\mathrm{c.u.c.}}\exp\left(\imath(\mathbf{Q}-\mathbf{q}'_{\mathrm{cyc}})\cdot\mathbf{d}-\imath\varphi_d\right) \tag*{(38)}

Now, the sums in Eq. (37), (38) are over a single chemical unit cell (c.u.c.). This can be rewritten, due to the periodicity of the lattice, as

IcalM=Φ∣−2.7fmag(Q)∣2(X⊥2+Y⊥2)(N(2π)3v0)⋅[∣F~+(Q)∣2∑Gδ(Q+qcyc′−G)+∣F~−(Q)∣2∑Gδ(Q−qcyc′−G)](39)\begin{aligned} I_{\mathrm{cal}}^M=\Phi\left|-2.7f_{\mathrm{mag}}(\mathbf{Q})\right|^2\left(\mathbf{X}_{\perp}^2+\mathbf{Y}_{\perp}^2\right)\left(N\frac{(2\pi)^3}{v_0}\right) \\ &\cdot\left[\left|\widetilde{F}^{+}(\mathbf{Q})\right|^2\sum_{\mathbf{G}}\delta(\mathbf{Q}+\mathbf{q}'_{\mathrm{cyc}}-\mathbf{G})+\left|\widetilde{F}^{-}(\mathbf{Q})\right|^2\sum_{\mathbf{G}}\delta(\mathbf{Q}-\mathbf{q}'_{\mathrm{cyc}}-\mathbf{G})\right] \tag*{(39)} \end{aligned}

where G\mathbf{G} are the reciprocal lattice vectors. Enforced by the δ\delta-functions, the F~±\widetilde{F}^{\pm} depend only on the reciprocal lattice vector G0(Q)\mathbf{G}_0(\mathbf{Q}) from which the incommensurate reflection ’originates’, i.e. Q=G0+qcyc′\mathbf{Q} = \mathbf{G}_0 + \mathbf{q}'_{\mathrm{cyc}} and F~±(Q)≡F~±(G0)\widetilde{F}^{\pm}(\mathbf{Q}) \equiv\widetilde{F}^{\pm}(\mathbf{G}_0), with

G0=H02πaea+K02πbeb+L02πcec(40)\mathbf{G}_0 = H_0 \frac{2\pi}{a}\mathbf{e}_a + K_0 \frac{2\pi}{b}\mathbf{e}_b + L_0 \frac{2\pi}{c}\mathbf{e}_c \tag*{(40)}

and H0H_0, K0K_0, L0L_0 integers.

Specializing to the (0KL)(0KL) scattering plane, F~±\widetilde{F}^{\pm} can be calculated explicitly by considering four types of pairs d=d′\mathbf{d} = \mathbf{d}', as well as four types of non-identical partners (we omit the aa-component of each vector),

d1−d3=d2−d4=b2eb,Δφ1,3=−π(41)\mathbf{d}_1 - \mathbf{d}_3 = \mathbf{d}_2 - \mathbf{d}_4 = \frac{b}{2}\mathbf{e}_b, \qquad\Delta\varphi_{1,3} = -\pi \tag*{(41)}
d2−d1=d4−d3=c2ec−b0eb,Δφ2,1=δ(42)\mathbf{d}_2 - \mathbf{d}_1 = \mathbf{d}_4 - \mathbf{d}_3 = \frac{c}{2}\mathbf{e}_c - b_0\mathbf{e}_b, \qquad\Delta\varphi_{2,1} = \delta \tag*{(42)}
d4−d1=c2ec−(b2+b0)eb,Δφ4,1=δ+π(43)\mathbf{d}_4 - \mathbf{d}_1 = \frac{c}{2}\mathbf{e}_c - \left(\frac{b}{2} + b_0\right)\mathbf{e}_b, \qquad\Delta\varphi_{4,1} = \delta+ \pi \tag*{(43)}
d2−d3=c2ec+(b2−b0)eb,Δφ2,3=δ−π(44)\mathbf{d}_2 - \mathbf{d}_3 = \frac{c}{2}\mathbf{e}_c + \left(\frac{b}{2} - b_0\right)\mathbf{e}_b, \qquad\Delta\varphi_{2,3} = \delta- \pi \tag*{(44)}

with b0b_0 the spacing, close to b/6b/6, between two layers in a covalently bonded bilayer (Figure E3). Due to the above definition of qcyc′\mathbf{q}'_{\mathrm{cyc}}, these phases are identical for the calculation of both F~±(G0)\widetilde{F}^{\pm}(\mathbf{G}_0). The parameter δ\delta by itself is either negative and positive, depending on the domain (A′,B′)(A', B').

Executing Eqs. (37), (38) in terms of this limited set of atom pairs,

∣F~δ±(G0)∣2=4[1−cos⁡(πK0)]⋅[1+cos⁡(πL0)cos⁡(−2πb0bK0±δ)](45)\left|\widetilde{F}^{\pm}_{\delta}(\mathbf{G}_0)\right|^2 = 4\left[1-\cos(\pi K_0)\right]\cdot\left[1+\cos(\pi L_0)\cos\left(-\frac{2\pi b_0}{b}K_0\pm\delta\right)\right] \tag*{(45)}

The total scattering intensity for two equally populated domains of the δ\delta angle is hence zero for K0=K_0= even and, for K0=K_0= odd,

IcalM=Φ(2.7fmag(Q))2(X⊥2+Y⊥2)(N(2π)3ν0)⋅12[∣F~δ+(G0)∣2+∣F~δ−(G0)∣2](46)I_{\mathrm{cal}}^M = \Phi\left(2.7f_{\mathrm{mag}}(\mathbf{Q})\right)^2\left(X_{\perp}^2+Y_{\perp}^2\right)\left(N\frac{(2\pi)^3}{\nu_0}\right)\cdot\frac{1}{2}\left[\left|\widetilde{F}^{+}_{\delta}(\mathbf{G}_0)\right|^2+\left|\widetilde{F}^{-}_{\delta}(\mathbf{G}_0)\right|^2\right] \tag*{(46)}
[⋯ ]=8(1+cos⁡(πL0)⋅cos⁡(2πb0bK0)⋅cos⁡(±δ))(47)\left[\cdots\right] = 8\left(1+\cos(\pi L_0)\cdot\cos\left(\frac{2\pi b_0}{b}K_0\right)\cdot\cos(\pm\delta)\right) \tag*{(47)}

independent of whether we are looking at the left / right reflection; i.e., the ±\pm sign in F~±\widetilde{F}^{\pm} is completely cancelled out. It is, finally, possible to simplify the magnetic moment projection as

X⊥2=X2−(X⋅Q^)2(48)X_{\perp}^2 = X^2 - \left(\mathbf{X}\cdot\hat{\mathbf{Q}}\right)^2 \tag*{(48)}
(X⊥2+Y⊥2)/X2=sin⁡2β+γ2cos⁡2β(49)\left(X_{\perp}^2+Y_{\perp}^2\right)/X^2 = \sin^2\beta+\gamma^2\cos^2\beta \tag*{(49)}

with γ2=Y2/X2\gamma^2 = Y^2/X^2, β=∠(Q^,ec)\beta= \angle(\hat{\mathbf{Q}}, \mathbf{e}_{c}), and we set eX=ec\mathbf{e}_{X} = \mathbf{e}_{c}, eY=eb\mathbf{e}_{Y} = \mathbf{e}_{b}.

Example: intensity ratio of two reflections

HHKKLLQaQ_aQbQ_bQcQ_c∣Q∣|\mathbf{Q}|2θ2\thetaL(θ)\mathcal{L}(\theta)fmagf_{\mathrm{mag}}β\beta
(r.l.u.)(Å−1^{-1})(deg.)(deg.)
090.20702.2280.3022.24851.8531.2720.82582.27
011.20700.2481.7631.78040.5051.5330.8857.99

Table E2. Calculation of expected intensity ratio of scattered neutron intensity at (0, 9, 0.207) and (0, 1, 1.207), based on magnetic structure model. The intensity ratio rr is calculated from Eq. (50), using these parameters. For convenience, we use L=1/sin⁡(2θ)\mathcal{L} = 1/\sin(2\theta) (neutron wavelength λ=1\lambda= 1) in the calculation of the Lorentz factor.

Specifically for two reflections Q1\mathbf{Q}_{1} and Q2\mathbf{Q}_{2} in reciprocal space, with Miller indices (0, 9, 0.207) and (0, 1, 1.207), the corresponding Miller indices for the G0\mathbf{G}_{0} positions are (090) and (011), respectively. According to Eq. (46) with δ=79∘\delta= 79^\circ from the refinement in § 6.6, the intensity ratio is

r=IcalM(Q1)IcalM(Q2)=L1L2⋅sin⁡2β1+γ2cos⁡2β1sin⁡2β2+γ2cos⁡2β2⋅fmag(Q1)2fmag(Q2)2⋅1+cos⁡(18πb0b)cos⁡δ1−cos⁡δ(50)r = \frac{I_{\mathrm{cal}}^{M}(\mathbf{Q}_{1})}{I_{\mathrm{cal}}^{M}(\mathbf{Q}_{2})} = \frac{\mathcal{L}_{1}}{\mathcal{L}_{2}} \cdot\frac{\sin^{2}\beta_{1}+\gamma^{2}\cos^{2}\beta_{1}}{\sin^{2}\beta_{2}+\gamma^{2}\cos^{2}\beta_{2}} \cdot\frac{f_{\mathrm{mag}}(\mathbf{Q}_{1})^{2}}{f_{\mathrm{mag}}(\mathbf{Q}_{2})^{2}} \cdot\frac{1+\cos\left(18\pi\frac{b_{0}}{b}\right)\cos\delta}{1-\cos\delta} \tag*{(50)}

where L=1/sin⁡(2θ)\mathcal{L} = 1/\sin(2\theta) (scattering angle 2θ2\theta) is the Lorentz factor, which corrects for the scattering geometry. Table E2 shows the numerical values for various steps in the calculation. From the observed ratio of peak intensities in Figure 2, r=0.77r = 0.77 and γ=Y/X≈0.97\gamma= Y/X \approx0.97 for these two reflections measured on sample A. This γ\gamma is somewhat larger than the result for sample B with full refinement in § 6.6, but sample A is a larger crystal, with anisotropic shape, where absorption correction is not applied. In particular, the observed intensities are expected to be larger along (0, 1, LL), close to transmission geometry. Such limitations of the data quality for sample A do not affect polarization analysis and the qualitative evolution of line scan intensities with temperature.

Observed magnetic structure factor and analytic model envelopes

Figure E6. Testing analytic model for the magnetic structure factor of the magnetic ground state of DyTe3. Observed magnetic structure factor, after correction for the Lorentz factor, for incommensurate reflections with L0=0L_0 = 0. The inset depicts the relevant region of momentum space in the (0KL)(0KL) scattering plane. The red curves (black dashed line) are envelopes to the data (is the center of the envelopes), bounding the data which oscillates according to Eq. (54). Blue and green are reflections with ±q′\pm\mathbf{q}', respectively. The oscillation amplitude rules out ∣cos⁡δ∣>0.4|\cos\delta| > 0.4, as discussed in section E4 D.

Structure factor for counter-propagating cycloids

Previously, a stack of counter-propagating helices was reported in γ\gamma-Li2_2IrO3_3 and ascribed to the presence of Kitaev-type anisotropic exchange interactions [80]. Analogously, we consider the structure factor for the helicity stack h=(+1,−1,+1,−1)\mathbf{h} = (+1,-1,+1,-1) as defined in section E4 A. In Eq. (35), δ=±πLcyc\delta= \pm\pi L_{\mathrm{cyc}} (δ=π±πLcyc\delta= \pi\pm\pi L_{\mathrm{cyc}}) can describe counter-propagating cycloids where the textures of two layers d=1,2d = 1,2 meet, at certain points, to create pairs of moments that are parallel to the bb-axis (parallel to the cc-axis). Recognizing that

m⊥=X⊥cos⁡(hdq′⋅rj+φd)+Y⊥sin⁡(hdq′⋅rj+φd)=(X⊥+ihdY⊥)exp⁡(iq′⋅rj+ihdφd)+(X⊥−ihdY⊥)exp⁡(−iq′⋅rj−ihdφd),(51)\begin{aligned} \mathbf{m}_{\perp} &= \mathbf{X}_{\perp}\cos(h_d\mathbf{q}'\cdot\mathbf{r}_j+\varphi_d)+\mathbf{Y}_{\perp}\sin(h_d\mathbf{q}'\cdot\mathbf{r}_j+\varphi_d) \\ &=(\mathbf{X}_{\perp}+ih_d\mathbf{Y}_{\perp})\exp(i\mathbf{q}'\cdot\mathbf{r}_j+ih_d\varphi_d) \\ &\quad+(\mathbf{X}_{\perp}-ih_d\mathbf{Y}_{\perp})\exp(-i\mathbf{q}'\cdot\mathbf{r}_j-ih_d\varphi_d), \tag*{(51)} \end{aligned}

algebra in line with section E4 B, for one domain of the δ\delta-angle, yields

Ical,±M=Φ(2.7fmag(Q))2(N(2π)3ν0)⋅4[1−cos⁡(πK0)]⋅[X⊥2+Y⊥2+(X⊥2−Y⊥2)cos⁡τ∓2X⊥⋅Y⊥⋅sin⁡τ](52)I^{M}_{\mathrm{cal},\pm}=\Phi\left(2.7f_{\mathrm{mag}}(Q)\right)^2\left(N\frac{(2\pi)^3}{\nu_0}\right)\cdot4\left[1-\cos(\pi K_0)\right]\cdot\left[\mathbf{X}_{\perp}^{2}+\mathbf{Y}_{\perp}^{2}+\left(\mathbf{X}_{\perp}^{2}-\mathbf{Y}_{\perp}^{2}\right)\cos\tau\mp2\mathbf{X}_{\perp}\cdot\mathbf{Y}_{\perp}\cdot\sin\tau\right] \tag*{(52)}
τ=2πb0bK0−πL0±δ(53)\tau=\frac{2\pi b_0}{b}K_0-\pi L_0\pm\delta \tag*{(53)}

Special cases can be considered. When Q\mathbf{Q} is (nearly) parallel to eb\mathbf{e}_b, Y⊥≈0\mathbf{Y}_{\perp}\approx0 and the co- and counter-propagating models deliver the same result (now including two δ\delta-domains, introducing a separate cos⁡δ\cos\delta factor):

Ical,±M∝(1−cos⁡(πK0))X⊥2⋅(1+cos⁡(πL0)cos⁡δcos⁡(2πb0bK0))(54)I^{M}_{\mathrm{cal},\pm}\propto(1-\cos(\pi K_0))\mathbf{X}_{\perp}^{2}\cdot\left(1+\cos(\pi L_0)\cos\delta\cos\left(\frac{2\pi b_0}{b}K_0\right)\right) \tag*{(54)}

This is expected from Eq. (51), where additional hdh_d’s appear only before Y⊥\mathbf{Y}_{\perp}. Figure E6 shows the intensities for L0=0L_0=0, as a function of K0K_0, with an upper and lower envelope function defined by the magnetic form factor. The maximum amplitude of the oscillation with KK, ∼40%\sim40\%, suggests that ∣cos⁡δ∣≤0.4|\cos\delta|\leq0.4, excluding δ=±πLcyc=±38∘\delta=\pm\pi L_{\mathrm{cyc}}=\pm38^\circ and δ=π±38∘\delta=\pi\pm38^\circ corresponding to counter-propagating cycloids with high symmetry (see above).

Analysis of nuclear elastic neutron scattering

For sample B, we collected 20 nuclear reflections, which contain 17 independent reflections. Observed and calculated nuclear structure factors are compared to determine the scale factor, a parameter for the extinction correction, and isotropic atomic displacement factors BisoB_{\mathrm{iso}}. These three parameters are later used in the magnetic structure analysis (§ 6.6).

For simplicity, we assumed that all the tellurium sites have the same BisoB_{\mathrm{iso}}. We performed a least-squares fit and found that the observed and calculated structure factors agree well; R(F)R(F) is 3.28 %. The values of BisoB_{\mathrm{iso}} for Dy and Te are determined to be 0.1(2) and 0.8(2), respectively. We also performed the same analysis after averaging the structure factors of the equivalent reflections, and obtained the internal R(F)R(F) value of 3.35 %.

For sample A, we collected 9 independent nuclear reflections. Observed and calculated nuclear structure factors are compared to determine the scale factor, while fractional coordinates of the atoms and BisoB_{\mathrm{iso}} are fixed at the values reported in [81]. We performed a least-squares fit and found that the observed and calculated structure factors agree well:

R(F)=7.4%R(F) = 7.4\% (Figure E8).

Results of magnetic structure analysis

For sample B, we collected 38 (70) Bragg reflections for the commensurate (incommensurate) magnetic reflections, all of which are independent. The polarized neutron scattering experiment in Figure E12 reveals that the magnetic moments corresponding to the commensurate component are parallel to the aa-axis. Assuming the volume fractions of two domains in Figure E3 to be equal, we performed a least-squares fit to the nonpolarized neutron scattering data from sample B and found that the magnitude of the commensurate component is 5.52(2) μB5.52(2)\,\mu_{\mathrm{B}}, with R(F)=5.78%R(F) = 5.78\%.

Neutron flipping ratio in sample A for two components of the magnetic order in DyTe3’s ground state

Figure E12. Neutron flipping ratio in sample A for two components of the magnetic order in DyTe3’s ground state. Here, the flipping ratio P=(INSF−ISF)/(INSF+ISF)P = (I_{\mathrm{NSF}} - I_{\mathrm{SF}})/(I_{\mathrm{NSF}} + I_{\mathrm{SF}}) is the normalized ratio of spin flip (SF) and non-spin flip (NSF) intensities for a magnetic reflection. P0P_{0} is the flipping ratio at the nuclear reflection (002) (Methods). In our geometry, SF intensity is dominated by the bb-component (the cc-component) of the magnetization at ω≈0∘\omega\approx0^{\circ} (at ω≈90∘\omega\approx90^{\circ}), although the structure and magnetic form factors have to be carefully taken into account when comparing intensities of various reflections (§ 6.6). a, The incommensurate magnetization component at qcyc\mathbf{q}_{\mathrm{cyc}} gives dominant SF scattering, with P/P0≡−1P/P_{0} \equiv-1 independent of the ω\omega angle. This implies presence of both mbm_{b} and mcm_{c} for the incommensurate reflection. b, The antiferromagnetic reflection qAFM\mathbf{q}_{\mathrm{AFM}} has dominant NSF scattering, which is again independent of ω\omega and consistent with magnetization component exclusively along the aa-axis; hence leading to P/P0≡+1P/P_{0} \equiv+1 for all reflections measured.

For the incommensurate magnetic component, polarized neutron scattering shows that the magnetic moments are confined in the bcbc-plane. Further considering the existence of the third higher-harmonic magnetic reflections, we assume an elliptic (distorted) cycloidal magnetic modulation, with principal axes parallel to the bb- and cc-axes, and moment amplitudes mbm_b and mcm_c common to all four Dy atoms in a chemical unit cell. We assume equal volume fraction of two domains for the phase shift δ\delta, described in section E4, and refine mbm_b, mcm_c, and δ\delta by least-squares analysis. This yields 3.93(3) μB3.93(3)\,\mu_{\mathrm{B}}, 6.61(3) μB6.61(3)\,\mu_{\mathrm{B}}, and ±79.2(5)\pm79.2(5) deg, respectively. The R(F)R(F) value is 8.26% (Figure E8) and combined, the total moment length at the Dy site is

mtot=ma2+max⁡(mb2,mc2)=8.61±0.1 μB/Dy(55)m_{\mathrm{tot}} = \sqrt{m_a^2 + \max\left(m_b^2,m_c^2\right)} = 8.61 \pm0.1\,\mu_{\mathrm{B}}/\mathrm{Dy} \tag*{(55)}

at T=2.2T = 2.2 K, which is about T/TN=0.6T/T_{\mathrm{N}} = 0.6, i.e. still rather close to the critical temperature; the ordered moment at absolute zero temperature is expected to be close to the local-ion value of 10 μB\mu_{\mathrm{B}} per dysprosium, also due to an underestimation in our calculation by neglecting higher harmonics contributions. The difference in value between mbm_b and mcm_c is consistent with the presence of third harmonic magnetic reflections; such distortions and higher harmonics have also been observed in insulating multiferroics [82].

Our model demonstrates that, at least, the incommensurate component involves both mbm_b and mcm_c components, of different amplitudes. Further experiments using spherical neutron polarimetry will be suitable to investigate, in more detail, the directions of the minor and major axes for the spin ellipsis in the bcbc plane.

The lattice constants a=c=4.302a = c = 4.302 Å, b=25.381b = 25.381 Å are determined from neutron scattering at low temperature. These values are close to the literature values for the orthorhombic, yet nearly tetragonal structure of DyTe3_3, a=c=4.296a = c = 4.296 Å, b=25.450b = 25.450 Å [81].

Our neutron measurements are not able to pick up minute atomic displacements due to the CDW, which can be seen in careful x-ray studies [75]; we used the fractional atomic positions of the average structure from Ref. [81] in our analysis of the nuclear and magnetic scattering.

Anisotropic electronic transport properties

Anisotropy of electrical transport properties in the ac basal plane of DyTe3

Figure E9. Anisotropy of electrical transport properties in acac basal plane of DyTe3_3. a,b, acac basal plane resistance ratio ρc/ρa\rho_c/\rho_a as a function of temperature for different samples. In Sample D, the assumed rectangular sample shape (c.f. panel c) is slightly – in particular 4% along the aa direction – adjusted to match ρc/ρa∼1\rho_c/\rho_a \sim1 at T>TCDWT > T_{\mathrm{CDW}}, c.f. Eq. (59). Both samples show a clear kink at the magnetic phase transition TN1T_{\mathrm{N1}}, followed by a broad plateau region, indicated by a red dashed line, with nearly constant ratio. c, Sample C mounted in Montgomery geometry. The orange rectangle depicts the assumed sample shape that is used to estimate the length ratio la/lc=1.266l_a/l_c = 1.266, where the white scale bar on the bottom right corresponds to 1 mm.

Angle-resolved photoemission (ARPES) studies combined with a tight binding (TB) model in Ref. [20] reveals that, in the family of the rear-earth tritellurides RTe3R\mathrm{Te}_3, the states at the Fermi level are mainly formed by the in-plane pxp_x and pzp_z orbitals of the Te-A ions in the Te square net, c.f. Figure 1a. Indeed, they are well separated by more than 1 eV from other bands, already indicating anisotropic bonding and transport properties. With a standard four-probe method and a modified Montgomery geometry Ru et al. determined both in-plane (ρac\rho_{ac}) and out-of-plane resistance (ρb\rho_b) of various RTe3R\mathrm{Te}_3 compounds [22]. They differ by at least one order of magnitude (ρb≫ρac\rho_b \gg\rho_{ac}), consistent with a metallic Te square net as well as covalently bonded RTeR\mathrm{Te} slabs.

To further extract the in-plane anisotropy ratio ρc/ρa\rho_c/\rho_a of DyTe3_3, we use the Montgomery technique and follow the method described in Ref. [83]. From temperature dependent resistance measurements Ra(T)R_a(T) and Rc(T)R_c(T), we calculate the shape of a hypothetical isotropic sample with dimensions La×Lb×LcL_a \times L_b \times L_c and same absolute resistance values Ra′R'_a and Rc′R'_c

L=LcLa=12[1πln⁡(RcRa)+(1πln⁡(RcRa))2+4](56)L = \frac{L_c}{L_a} = \frac{1}{2}\left[\frac{1}{\pi}\ln\left(\frac{R_c}{R_a}\right) + \sqrt{\left(\frac{1}{\pi}\ln\left(\frac{R_c}{R_a}\right)\right)^2 + 4}\right] \tag*{(56)}

Based on published data from Ref. [22] and our sample geometry, a thin exfoliated flake, we estimated that the calculation of the anisotropy ratio ρc/ρa\rho_c/\rho_a is possible in the thin-layer limit. Following Ref. [83] and using the real sample dimensions la×lb×lcl_a \times l_b \times l_c this leads to

ρa=π8lblclaL−1Rasinh⁡(πL)(57)\rho_a = \frac{\pi}{8}\frac{l_b l_c}{l_a}L^{-1}R_a\sinh(\pi L) \tag*{(57)}
ρc=π8lblalcLRasinh⁡(πL)(58)\rho_c = \frac{\pi}{8}\frac{l_b l_a}{l_c}L R_a\sinh(\pi L) \tag*{(58)}

and finally to

ρcρa=(lalcL)2(59)\frac{\rho_c}{\rho_a} = \left(\frac{l_a}{l_c}L\right)^2 \tag*{(59)}

There is some ambiguity in the interpretation of resistance anisotropy changes at TNT_N, which can be ascribed either to opening of a partial charge gap in the ordered state, or to fluctuations in the paramagnetic regime and their suppression below TNT_N. In DyTe3_3, the sign of the observed change in ρc/ρa\rho_c/\rho_a implies that, when moving from the paramagnetic into the ordered regime, the resistance along the cc-axis becomes larger than the resistance along the aa-axis. This result can be neatly explained by attributing partial gap opening to qAFM=(0,b∗,0.5c∗)\mathbf{q}_{\mathrm{AFM}} = (0,b^*,0.5c^*), qcyc=(0,b∗,0.207c∗)\mathbf{q}_{\mathrm{cyc}} = (0,b^*,0.207c^*), while unidirectional fluctuations should enhance ρc/ρa\rho_c/\rho_a above TNT_N, and suppress it below TNT_N.

References

  1. [1]B. Huang, G. Clark, E. Navarro-Moratalla, D. Klein, R. Cheng, K. Seyler, E. S. D. Zhong, M. McGuire, D. Cobden, W. Yao, D. Xiao, P. Jarillo-Herrero, and X. Xu, Layer-dependent ferromagnetism in a van der Waals crystal down to the monolayer limit, Nature 546, 270 (2017).DOI
  2. [2]Gong, C., Li, L., Li, Z., Ji, H., Stern, A., Xia, Y., Cao, T., Bao, W., Wang, C., Wang, Y., Qiu, Z., Cava, R., Louie, S., Xia, J., & Zhang, X. (n.d.). Discovery of intrinsic ferromagnetism in.
  3. [3]K. Burch, D. Mandrus, and J.-G. Park, Magnetism in two-dimensional van der Waals materials, Nature 563, 47 (2018).DOI
  4. [4]C. Gong and X. Zhang, Two-dimensional magnetic crystals and emergent heterostructure devices, Science 363, aav4450 (2019).DOI
  5. [5]D. Amoroso, P. Barone, and S. Picozzi, Spontaneous skyrmionic lattice from anisotropic symmetric exchange in a Ni-halide monolayer, Nature Communications 11, 5784 (2020).DOI
  6. [6]K. Shimizu, S. Okumura, Y. Kato, and Y. Motome, Spin moiré engineering of topological magnetism and emergent electromagnetic fields, Physical Review B 103, 184421 (2021).DOI
  7. [7]N. Jiang, Y. Nii, H. Arisawa, E. Saitoh, and Y. Onose, Electric current control of spin helicity in an itinerant helimagnet, Nature Communications 11, 1601 (2020).DOI
  8. [8]J. Ohe and Y. Onose, Chirality control of the spin structure in monoaxial helimagnets by charge current, Applied Physics Letters 118, 042407 (2021).DOI
  9. [9]H. Masuda, T. Seki, J. Ohe, Y. Nii, K. Takanashi, and Y. Onose, Chirality-dependent spin current generation in a helimagnet: zero-field probe of chirality, arXiv:2212.10980 (2022).arxiv.org/abs/2212.10980
  10. [10]Y. Wang, X. Xu, X. Zhao, W. Ji, Q. Cao, S. Li, and Y. Li, Switchable half-metallicity in A-type antiferromagnetic NiI₂ bilayer coupled with ferroelectric In₂Se₃, npj Computational Materials 8, 218 (2022).DOI
  11. [11]F. Schmitt, P. Kirchmann, U. Bovensiepen, R. Moore, J.-H. Chu, D. Lu, L. Rettig, M. Wolf, I. Fisher, and Z.-X. Shen, Ultrafast electron dynamics in the charge density wave material TbTe₃, New Journal of Physics 13, 063022 (2011).DOI
  12. [12]A. Kogar, A. Z. ad P.E. Dolgirev, X. Shen, J. Straquadine, Y.-Q. Bie, X. Wang, T. Rohwer, I.-C. Tung, Y. Yang, R. Li, J. Yang, S. Weathersby, S. Park, M. Kozina, E. Sie, H. Wen, P. Jarillo-Herrero, I. Fisher, X. Wang, and N. Gedik, Light-induced charge density wave in LaTe₃, Nature Physics 16, 159 (2020).DOI
  13. [13]P. Dolgirev, A. Rozhkov, A. Zong, A. Kogar, N. Gedik, and B. Fine, Amplitude dynamics of the charge density wave in LaTe₃: Theoretical description of pump-probe experiments, Physical Review B 101, 054203 (2020).DOI
  14. [14]I. Gonzalez-Vallejo, V. Jacques, D. Boschetto, G. Rizza, A. Hadj-Azzem, J. Faure, and D. L. Bolloc’h, Time-resolved structural dynamics of the out-of-equilibrium charge density wave phase transition in GdTe₃, Structural Dynamics 9, 014502 (2022).DOI
  15. [15]Y. Wang, I. Petrides, G. McNamara, M. Hosen, S. Lei, Y.-C. Wu, J. Hart, H. Lv, J. Yan, D. Xiao, J. Cha, P. Narang, L. Schoop, and K. Burch, Axial Higgs mode detected by quantum pathway interference in RTe₃, Nature 606, 896 (2022).DOI
  16. [16]S. Lei, J. Lin, Y. Jia, M. Gray, A. Topp, G. Farahi, S. Klemenz, T. Gao, F. Rodolakis, J. McChesney, C. Ast, A. Yazdani, K. Burch, S. Wu, N. Ong, and L. Schoop, High mobility in a van der Waals layered antiferromagnetic metal, Science Advances 6, eaay6407 (2020).DOI
  17. [17]Y. Che, P. Wang, M. Wu, J. Ma, S. Wen, X. Wu, G. Li, Y. Zhao, K. Wang, L. Zhang, L. Huang, W. Li, and M. Huang, Raman spectra and dimensional effect on the charge density wave transition in GdTe₃, Applied Physics Letters 115, 151905 (2019).DOI
  18. [18]C. Malliakas and M. Kanatzidis, Divergence in the Behavior of the Charge Density Wave in RETe₃ (RE= Rare-Earth element) with Temperature and RE Element, Journal of the American Chemical Society 128, 12612 (2006).pubmed.ncbi.nlm.nih.gov/17002331
  19. [19]J. Laverock, S. B. Dugdale, Z. Major, M. A. Alam, N. Ru, I. R. Fisher, G. Santi, and E. Bruno, Fermi surface nesting and charge-density wave formation in rare-earth tritellurides, Physical Review B 71, 085114 (2005).DOI
  20. [20]V. Brouet, W. L. Yang, X. J. Zhou, Z. Hussain, R. G. Moore, R. He, D. H. Lu, Z. X. Shen, J. Laverock, S. B. Dugdale, N. Ru, and I. R. Fisher, Angle-resolved photoemission study of the evolution of band structure and charge density wave properties in RTe₃ (R = Y, La, Ce, Sm, Gd, Tb, and Dy), Physical Review B 77, 235104 (2008).DOI
  21. [21]A. Chikina, H. Lund, M. Bianchia, D. Curcio, K. Dalgaard, M. Bremholm, S. Lei, R. Singh, L. Schoop, and P. Hofmann, Charge density wave-generated Fermi surfaces in NdTe₃, (2022).DOI
  22. [22]Ru, N., Condron, C., Margulis, G., Shin, K., Laverock, J., Dugdale, S., Toney, M., & Fisher, I. (2008). Effect of chemical pressure on the charge density wave transition in rare-earth tritellurides RTe₃. Physical Review B, 77, 035114.
  23. [23]D. A. Zocco, Interplay of Superconductivity, Magnetism, and Density Waves in Rare-Earth Tritellurides and Iron-Based Superconducting Materials, Ph.D. thesis, University of California, San Diego (2011).
  24. [24]D. A. Zocco, J. J. Hamlin, K. Grube, J.-H. Chu, H.-H. Kuo, I. R. Fisher, and M. B. Maple, Pressure dependence of the charge-density-wave and superconducting states in GdTe₃, TbTe₃, and DyTe₃, Physical Review B 91, 205114 (2015).DOI
  25. [25]S. Sarkar, J. Bhattacharya, P. Sadhukhan, D. Curcio, R. Dutt, V. Singh, M. Bianchi, A. Pari- ari, S. Roy, P. Mandal, T. Das, P. Hofmann, A. Chakrabarti, and S. Barman, Charge density wave induced nodal lines in LaTe3, Nature Communications 14, 3628 (2023).DOI
  26. [26]Y. Iyeiri, T. Okumura, C. Michioka, and K. Suzuki, Magnetic properties of rare-earth metal tritellurides RTe3 (R = Ce, Pr, Nd, Gd, Dy), Physical Review B 67, 144417 (2003).DOI
  27. [27]F. Pfuner, S. Gvasaliya, O. Zaharko, L. Keller, J. Mesot, V. Pomjakushin, J.-H. Chu, I. Fisher, and L. Degiorgi, Incommensurate magnetic order in TbTe3, Journal of Physics: Condensed Matter 24, 036001 (2011).DOI
  28. [28]Z. Yang, A. Drew, S. van Smaalen, N. van Well, F. Pratt, G. Stenning, A. Karim, and K. Rabia, Multiple magnetic-phase transitions and critical behavior of charge-density wave compound TbTe3, Journal of Physics: Condensed Matter 32, 305801 (2020).DOI
  29. [29]Q. Guo, D. Bao, L. Zhao, and S. Ebisu, Novel magnetic behavior of antiferromagnetic GdTe3 induced by magnetic field, Physica B: Condensed Matter 617, 413153 (2021).DOI
  30. [30]O. Volkova, A. Hadj-Azzem, G. Remenyi, J. Lorenzo, P. Monceau, A. Sinchenko, and A. Vasiliev, Magnetic Phase Diagram of van der Waals Antiferromagnet TbTe3, Materials 15, 8772 (2022).DOI
  31. [31]S. Chillal, E. Schierle, E. Weschke, F. Yokaichiya, J.-U. Hoffmann, O. S. Volkova, A. N. Vasiliev, A. Sinchenko, P. Lejay, A. Hadj-Azzem, P. Monceau, and B. Lake, Strongly coupled charge, orbital, and spin order in TbTe3, Physical Review B 102, 241110(R) (2020).DOI
  32. [32]K. Y. Shin, V. Brouet, N. Ru, Z. X. Shen, and I. R. Fisher, Electronic structure and charge-density wave formation in LaTe1.95 and CeTe2.00, Physical Review B 72, 085132 (2005).DOI
  33. [33]J. Straquadine, M. S. Ikeda, and I. Fisher, Evidence for Realignment of the Charge Density Wave State in ErTe3 and TmTe3 under Uniaxial Stress via Elastocaloric and Elastoresistivity Measurements, Physical Review X 12, 021046 (2022).DOI
  34. [34]R. Dickinson and L. Pauling, The crystal structure of molybdenite, Journal of the American Chemical Society 45, 1466 (1923).DOI
  35. [35]S. Manzeli, D. Ovchinnikov, D. Pasquier, O. Yazyev, and A. Kis, 2D transition metal dichalcogenides, Nature Reviews Materials 2, 17033 (2017).DOI
  36. [36]B. Stärk, P. Krüger, and J. Pollmann, Magnetic anisotropy of thin Co and Ni films on diamond surfaces, Physical Review B 84, 195316 (2011).DOI
  37. [37]E. Torun, H. Sahin, C. Bacaksiz, R. T. Senger, and F. M. Peeters, Tuning the magnetic anisotropy in single-layer crystal structures, Physical Review B 92, 104407 (2015).DOI
  38. [38]P. Gambardella, S. Stepanow, A. Dmitriev, J. Honolka, F. de Groot, M. Lingenfelder, S. Gupta, D. Sarma, P. Bencok, S. Stanescu, S. Clair, S. Pons, N. Lin, A. Seitsonen, H. Brune, J. Barth, and K. Kern, Supramolecular control of the magnetic anisotropy in two-dimensional high-spin Fe arrays at a metal interface, Nature Materials 8, 189 (2009).DOI
  39. [39]T. Honda, J. S. White, A. B. Harris, L. C. Chapon, A. Fennell, B. Roessli, O. Zaharko, Y. Murakami, M. Kenzelmann, and T. Kimura, Coupled multiferroic domain switching in the canted conical spin spiral system Mn₂GeO₄, Nature Communications 8, 15457 (2017).DOI
  40. [40]S. Riberolles, T. Trevisan, B. Kuthanazhi, T. Heitmann, F. Ye, D. Johnston, S. Bud’ko, D. Ryan, P. Canfield, A. Kreyssig, A. Vishwanath, R. McQueeney, L.-L. Wang, P. Orth, and B. Ueland, Magnetic crystalline-symmetry-protected axion electrodynamics and field-tunable unpinned Dirac cones in EuIn₂As₂, Nature Communications 12, 999 (2021).DOI
  41. [41]S. Lei, V. Duppel, J. Lippmann, J. Nuss, B. Lotsch, and L. Schoop, Charge Density Waves and Magnetism in Topological Semimetal Candidates GdSbxTe₂−x−δ, Advanced Quantum Technologies 2, 1900045 (2019).DOI
  42. [42]S. Lei, A. Saltzman, and L. Schoop, Complex magnetic phases enriched by charge density waves in the topological semimetals GdSbxTe₂−x−δ, Physical Review B 103, 134418 (2021).DOI
  43. [43]Grüner, G. (1988). The dynamics of charge-density waves. Reviews of Modern Physics, 60, 1129.
  44. [44]A. A. Sinchenko, P. Lejay, and P. Monceau, Sliding charge-density wave in two-dimensional rare-earth tellurides, Physical Review B 85, 241104 (2012).DOI
  45. [45]T. Moriya, Anisotropic Superexchange Interaction and Weak Ferromagnetism, Physical Review 120, 91 (1960).DOI
  46. [46]D. Ghader, B. Jabakhanji, and A. Stroppa, Whirling interlayer fields as a source of stable topological order in Moiré CrI₃, Communications Physics 5, 192 (2022).DOI
  47. [47]G. Volovik, Linear momentum in ferromagnets, Journal of Physics C: Solid State Physics 20, L83 (1987).DOI
  48. [48]Y. Tokura and N. Nagaosa, Nonreciprocal responses from non-centrosymmetric quantum materials, Nature Communications 9, 3740 (2018).DOI
  49. [49]J. A. Osborn, Demagnetizing Factors of the General Ellipsoid, Physical Review 67, 351 (1945).DOI
  50. [50]A. Scheie, PyCrystalField: Software for Calculation, Analysis and Fitting of Crystal Electric Field Hamiltonians, Journal of Applied Crystallography 54, 356 (2021).DOI
  51. [51]Y. Gao, Q. Yin, Q. Wang, Z. Li, J. Cai, T. Zhao, H. Lei, S. Wang, Y. Zhang, and B. Shen, Spon- taneous (Anti)meron Chains in the Domain Walls of van der Waals Ferromagnetic Fe₅−ₓGeTe₂, Advanced Materials 32, 2005228 (2020).DOI
  52. [52]T. T. Ly, J. Park, K. Kim, H.-B. Ahn, N. J. Lee, K. Kim, T.-E. Park, G. Duvjir, N. H. Lam, K. Jang, C.-Y. You, Y. Jo, S. K. Kim, C. Lee, S. Kim, and J. Kim, Direct Observation of Fe-Ge Ordering in Fe₅−ₓGeTe₂ Crystals and Resultant Helimagnetism, Advanced Functional Materials 31, 2009758 (2021).DOI
  53. [53]A. F. May, C. A. Bridges, and M. A. McGuire, Physical properties and thermal stability of Fe₅−ₓGeTe₂ single crystals, Physical Review Materials 3, 104401 (2019).DOI
  54. [54]M. Baenitz, M. M. Piva, S. Luther, J. Sichelschmidt, K. M. Ranjith, H. Dawczak-Dębicki, M. O. Ajeesh, S.-J. Kim, G. Siemann, C. Bigi, P. Manuel, D. Khalyavin, D. A. Sokolov, P. Mokhtari, H. Zhang, H. Yasuoka, P. D. C. King, G. Vinai, V. Polewczyk, P. Torelli, J. Wosnitza, U. Burkhardt, B. Schmidt, H. Rosner, S. Wirth, H. Kühne, M. Nicklas, and M. Schmidt, Planar triangular S = 3/2 magnet AgCrSe₂: Magnetic frustration, short range correlations, and field-tuned anisotropic cycloidal magnetic order, Physical Review B 104, 134410 (2021).DOI
  55. [55]U. K. Gautam, R. Seshadri, S. Vasudevan, and A. Maignan, Magnetic and transport properties, and electronic structure of the layered chalcogenide AgCrSe₂, Solid state communications 122, 607 (2002).DOI
  56. [56]T. Kurumaji, S. Seki, S. Ishiwata, H. Murakawa, Y. Kaneko, and Y. Tokura, Magnetoelectric responses induced by domain rearrangement and spin structural change in triangular-lattice helimagnets NiI₂ and CoI₂, Physical Review B 87, 014429 (2013).DOI
  57. [57]Lebedev, D., Gish, J. T., Garvey, E. S., Stanev, T. K., Choi, J., Georgopoulos, L., Song, T. W., Park, H. Y., Watanabe, K., Taniguchi, T., Stern, N. P., Sangwan, V. K., & Hersam, M. C. (2023). Electrical Interrogation of Thickness-Dependent Multiferroic Phase Transitions in the 2D Antiferromagnetic Semiconductor NiI$_2$. Advanced Functional Materials, 33, 2212568.
  58. [58]A. Adam, D. Billerey, C. Terrier, R. Mainard, L. Regnault, J. Rossat-Mignod, and P. Mériel, Neutron diffraction study of the commensurate and incommensurate magnetic structures of NiBr₂, Solid State Communications 35, 1 (1980).DOI
  59. [59]Y. Tokunaga, D. Okuyama, T. Kurumaji, T. Arima, H. Nakao, Y. Murakami, Y. Taguchi, and Y. Tokura, Multiferroicity in NiBr₂ with long-wavelength cycloidal spin structure on a triangular lattice, Physical Review B 84, 060406 (2011).DOI
  60. [60]C. R. Ronda, G. J. Arends, and C. Haas, Photoconductivity of the nickel dihalides and the nature of the energy gap, Physical Review B 35, 4038 (1987).pubmed.ncbi.nlm.nih.gov/9941930
  61. [61]Kurumaji, T., Seki, S., Ishiwata, S., Murakawa, H., Tokunaga, Y., Kaneko, Y., & Tokura, Y. (2011). Magnetic-Field Induced Competition of Two Multiferroic Orders in a Triangular-Lattice Helimagnet $\mathrm{MnI}_{2}$. Physical Review Letters, 106, 167206.
  62. [62]N. J. Ghimire, M. A. McGuire, D. S. Parker, B. Sipos, S. Tang, J.-Q. Yan, B. C. Sales, and D. Mandrus, Magnetic phase transition in single crystals of the chiral helimagnet Cr₁/₃NbS₂, Physical Review B 87, 104403 (2013).DOI
  63. [63]K. Lu, A. Murzabekova, S. Shim, J. Park, S. Kim, L. Kish, Y. Wu, L. DeBeer-Schmitt, A. A. Aczel, A. Schleife, N. Mason, F. Mahmood, and G. J. MacDougall, Understanding the Anomalous Hall effect in Co₁/₃NbS₂ from crystal and magnetic structures (2022), arXiv:2212.14762 [cond-mat.mtrl-sci].arXiv
  64. [64]G. Tenasini, E. Martino, N. Ubrig, N. J. Ghimire, H. Berger, O. Zaharko, F. Wu, J. F. Mitchell, I. Martin, L. Forró, and A. F. Morpurgo, Giant anomalous Hall effect in quasi-two-dimensional layered antiferromagnet Co₁/₃NbS₂, Physical Review Research 2, 023051 (2020).DOI
  65. [65]H. Takagi, R. Takagi, S. Minami, T. Nomoto, K. Ohishi, M.-T. Suzuki, Y. Yanagi, M. Hirayama, N. Khanh, K. Karube, H. Saito, D. Hashizume, R. Kiyanagi, Y. Tokura, R. Arita, T. Nakajima, and S. Seki, Spontaneous topological Hall effect induced by non-coplanar antiferromagnetic order in intercalated van der Waals materials, Nature Physics 19, 961 (2023).DOI
  66. [66]Kousaka, Y., Ogura, T., Zhang, J., Miao, P., Lee, S., Torii, S., Kamiyama, T., Campo, J., Inoue, K., & Akimitsu, J. (2016). Long Periodic Helimagnetic Ordering in $\mathrm{CrM}_{3}\mathrm{S}_{6}$ ($M = \mathrm{Nb}$ and $\mathrm{Ta}$). Journal of Physics: Conference Series, 746, 012061.
  67. [67]Miyadai, T., Kikuchi, K., Kondo, H., Sakka, S., Arai, M., & Ishikawa, Y. (1983). Magnetic Properties of $\mathrm{Cr}_{1/3}\mathrm{NbS}_{2}$. Journal of the Physical Society of Japan, 52, 1394.
  68. [68]L. Wang, N. Chepiga, D.-K. Ki, L. Li, F. Li, W. Zhu, Y. Kato, O. S. Ovchinnikova, F. Mila, I. Martin, D. Mandrus, and A. F. Morpurgo, Controlling the Topological Sector of Magnetic Solitons in Exfoliated Cr₁/₃NbS₂ Crystals, Physical Revie Letters 118, 257203 (2017).DOI
  69. [69]Obeysekera, D., Gamage, K., Gao, Y., Cheong, S.-. w ., & Yang, J. (2021). The Magneto-Transport Properties of $\mathrm{Cr}_{1/3}\mathrm{TaS}_{2}$ with Chiral Magnetic Solitons. Advanced Electronic Materials, 7, 2100424.
  70. [70]C. Zhang, J. Zhang, C. Liu, S. Zhang, Y. Yuan, P. Li, Y. Wen, Z. Jiang, B. Zhou, Y. Lei, D. Zheng, C. Song, Z. Hou, W. Mi, U. Schwingenschlögl, A. Manchon, Z. Q. Qiu, H. N. Alshareef, Y. Peng, and X.-X. Zhang, Chiral Helimagnetism and One-Dimensional Magnetic Solitons in a Cr-Intercalated Transition Metal Dichalcogenide, Advanced Materials 33, 2101131 (2021).DOI
  71. [71]C.-H. Zhang, H. Algaidi, P. Li, Y. Yuan, and X.-X. Zhang, Magnetic soliton confinement and discretization effects in Cr₁/₃TaS₂ nanoflakes, Rare Metals 41, 3005 (2022).DOI
  72. [72]Tables of Form Factors. (2023). Tables of Form Factors. https://www.ill.eu/sites/ccsl/ffacts/ill.eu/sites/ccsl/ffacts
  73. [73]Squires, G. L. (2012). Introduction to the Theory of Thermal Neutron Scattering.
  74. [74]Slovyanskikh, V., Kuznetsov, N., & Gracheva, N. (1985). The Dy-U-Te system. Russian Journal of Inorganic Chemistry, 30, 1666.
  75. [75]C. Malliakas, S. J. L. Billinge, H. J. Kim, and M. G. Kanatzidis, Square Nets of Tellurium: Rare-Earth Dependent Variation in the Charge-Density Wave of RETe₃ (RE = Rare-Earth Element), Journal of the American Chemical Society 127, 6510 (2005).DOI
  76. [76]M. I. Aroyo, A. Kirov, C. Capillas, J. M. Perez-Mato, and H. Wondratschek, Bilbao Crystallographic Server II: Representations of crystallographic point groups and space groups, Acta Crystallographica A62, 115 (2006).pubmed.ncbi.nlm.nih.gov/16489249
  77. [77]Kenzelmann, M., Harris, A. B., Aharony, A., Entin-Wohlman, O., Yildirim, T., Huang, Q., Park, S., Lawes, G., Broholm, C., Rogado, N., Cava, R. J., Kim, K. H., Jorge, G., & Ramirez, A. P. (2006). Field dependence of magnetic ordering in Kagomé-staircase compound Ni$_3$V$_2$O$_8$. Phys. Rev. B, 74, 014429.
  78. [78]Brown, P., Chattopadhyay, T., Forsyth, J., & Nunez, V. (1991). Magnetic phase transitions of MnWO$_4$ studied by the use of neutron diffraction. J. Phys.: Condens. Matter, 48, 4281.
  79. [79]G. Lautenschläger, H. Weitzel, T. Vogt, R. Hock, A. Böhm, M. Bonnet, and H. Fuess, Magnetic phase transitions of MnWO₄ studied by the use of neutron diffraction, Phys. Rev. B 48, 6087 (1993).pubmed.ncbi.nlm.nih.gov/10009145
  80. [80]A. Biffin, R. D. Johnson, I. Kimchi, R. Morris, A. Bombardi, J. G. Analytis, A. Vishwanath, and R. Coldea, Noncoplanar and Counterrotating Incommensurate Magnetic Order Stabilized by Kitaev Interactions in γ−Li₂IrO₃, Phys. Rev. Lett. 113, 197201 (2014).pubmed.ncbi.nlm.nih.gov/25415919
  81. [81]Pardo, M. P., & Flahaut, J. (1967). Les tellurures superieurs des elements des terres rares, de formules L$_2$Te$_5$ et LTe$_3$. Bulletin de la Société Chimique de France, 3658.
  82. [82]M. Kenzelmann, A. B. Harris, S. Jonas, C. Broholm, J. Schefer, S. B. Kim, C. L. Zhang, S.-W. Cheong, O. P. Vajk, and J. W. Lynn, Magnetic Inversion Symmetry Breaking and Ferroelectricity in TbMnO3, Phys. Rev. Lett. 95, 087206 (2005).DOI
  83. [83]C. Dos Santos, A. De Campos, M. Da Luz, B. White, J. Neumeier, B. De Lima, and C. Shigue, Procedure for measuring electrical resistivity of anisotropic materials: A revision of the Montgomery method, Journal of Applied Physics 110 (2011).DOI

Paper details

vlnc:W_953ggzwg
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