betaFeedback

WorksPhysics and Astronomy

Spontaneous Anomalous Hall Effect in Two-Dimensional Altermagnets

Abstract

The anomalous Hall effect (AHE) is an efficient tool for detecting the Néel vector in collinear compensated magnets with spin-split bands, known as altermagnets (AMs). Here, we establish design principles for obtaining non-zero anomalous Hall conductivity in the recently proposed two-dimensional (2D) AMs using spin and magnetic group symmetry analysis. We show that only two of the seven nontrivial spin layer groups exhibit an unconventional in-plane AHE in which the Néel vector lies within the plane of the Hall current. Through first-principles simulations on bilayers of MnPSe3_3 and MnSe, we demonstrate the validity of our group theoretic framework for obtaining AHE with dd and ii-wave altermagnetic orders, depending on the stacking of the bilayers. We find that the spin group symmetry is successful in determining the linear and cubic dependence of anomalous Hall conductivity in Néel vector space, although AHE is a relativistic effect. This work shows that the AHE in 2D AMs can probe the altermagnetic order and Néel vector reversal, thereby facilitating the miniaturization of altermagnetic spintronics.

Introduction

Altermagnetism (AM) has emerged as a novel class of collinear magnetism characterized by time-reversal symmetry (TT) breaking in momentum space along with compensated magnetic order in real space [1, 2, 3, 4, 5]. Unlike conventional antiferromagnets (AFMs), where the opposite spin sublattices are connected by the inversion (PP) and/or translation (τ\tau) operations, in AMs they are connected by mirror-rotation symmetries. Hence, AMs show eV-scale nonrelativistic spin splittings [3], leading to highly spin-polarized currents, which are characteristic of ferromagnets (FMs). Since there is no net magnetization, AMs also show ultrafast switching dynamics and resilience towards stray fields, similar to that shown by AFMs [6, 7, 8, 9]. This combination of properties of both FMs and AFMs in a single material makes AMs interesting not only for fundamental research, but also for spintronics-based applications by allowing for facile control and detection of spin states using different means, including electric or optical fields. Broken TT-symmetry effects in AMs are experimentally detected through angle-resolved photoemission spectroscopy [10, 11], spin-to-charge interconversion [12, 13], and the anomalous Hall effect (AHE) [14]. To date, the experimental observation of altermagnetic effects is still limited to three-dimensional (3D) materials, i.e., MnTe [15], Mn5_5Si3_3 [16], CrSb [17], and RuO2_2 [12].

The discovery of magnetic ordering in atomically thin materials opens up new possibilities for miniaturizing devices to the two-dimensional (2D) limit. However, achieving AM in 2D systems is difficult due to additional symmetry constraints. In 2D layers, the electronic bands are dispersionless along the out-of-plane direction [18–21]. As a result, in the 2D limit, the two-fold rotation along the zz-axis (C2zC_{2}^{z}) and horizontal mirror (MzM_{z}) symmetry transform k=(kx,ky)\boldsymbol{k}=(k_{x},k_{y}) like PP and τ\tau, respectively. Therefore, for altermagnetism to emerge, the opposite spin sublattices must not be connected by MzM_{z} and/or C2zC_{2}^{z} [19]. In spite of these exacting requirements, AM in 2D materials was predicted theoretically using a high-throughput computational approach (RuF4_4, FeBr2_2 [22, 23]), bilayer stackings (CrSBr, MnBi2_2Te4_4 [18, 20]) and twisting bilayers [21, 24, 25]. The nonrelativistic understanding of 2D altermagnetism was recently established in Refs. [18–25]. However, unlike 3D bulk AMs, a comprehensive analysis of the magnetotransport effects (such as anomalous Hall and anomalous Nernst effects), and novel Néel vector detection methods for 2D AMs is still lacking.

Our work based on symmetry analysis and first-principles simulations elucidates the stringent symmetry requirement for observing AHE in the 2D AMs. We show an unconventional periodic dependence of AHE on the Néel-vector space, where the Néel vector lies in the plane of the Hall current. This is in contrast to the conventional FM-like Hall response, where Hall current is perpendicular to the magnetization [26, 27]. By performing density functional theory (DFT) simulations on bilayers of MnPSe3_3 and MnSe, which are PTPT-symmetric as monolayers, we achieved dd- and ii-wave altermagnetism with an in-plane anomalous Hall response as large as for the well-known bulk AMs—MnTe [15, 28] and Mn5_5Si3_3 [16, 29]. Furthermore, we reveal the unique relationship of the AHE with the spin group symmetries and nonrelativistic spin-degenerate nodal lines. Moreover, the in-plane AHE is shown to be an efficient tool to probe the altermagnetic order and 180∘180^\circ Néel vector reversal in 2D AMs.

Calculation Methods

DFT calculations have been carried out using the projector-augmented wave method [30], as implemented in the VASP package [31]. The Perdew-Burke-Ernzerhof (PBE) functional within the generalized-gradient approximation, along with a Hubbard U correction, are employed to accurately describe electronic interactions. Following Refs. [33, 34], an effective U value of 3.0 eV was applied to the Mn-dd orbitals using the approach of Dudarev et al. [35]. The Grimme-D3 scheme was used to account for the vdW interactions [36].

PPCnzC_n^zCnxC_n^xMzM_zMxM_xS4,6zS_{4,6}^zS4,6xS_{4,6}^xTTPTPTTC2zTC_2^zTC2xTC_2^xTMzTM_zTMxTM_xTC2,3,4,6x,zTC_{2,3,4,6}^{x,z}TS4,6x,zTS_{4,6}^{x,z}
σxy\sigma_{xy}✓✓✗✓✗✓✗✗✗✗✓✗✓✗✗
σxz\sigma_{xz}✓✗✗✗✗✗✗✗✗✓✓✓✓✗✗
σyz\sigma_{yz}✓✗✓✗✓✗✓✗✗✓✗✓✗✗✗

Table 1. Symmetry constraints on the components of the anomalous Hall conductivity tensor imposed by the magnetic point group operations. ✓ and ✗ denotes symmetry allowed and forbidden components, respectively. The zz (xx) denotes the out-of-plane (in-plane) direction.

The AHE can be described by an order parameter called the Hall vector (σH\boldsymbol{\sigma}_{\mathrm{H}}), with the anomalous Hall current density given by J=σH×E\boldsymbol{J}=\boldsymbol{\sigma}_{\mathrm{H}}\times\boldsymbol{E}, where E\boldsymbol{E} is the electric field. The Hall vector is defined as: σH=(σzy,σxz,σyx)\boldsymbol{\sigma}_{\mathrm{H}}=(\sigma_{zy},\sigma_{xz},\sigma_{yx}) [37], with σαβ\sigma_{\alpha\beta} for α≠β\alpha\ne\beta being the antisymmetric components of the conductivity tensor.

At the microscopic level, the intrinsic contribution to anomalous Hall conductivity (AHC) is expressed as

σαβ=−e2ℏϵαβγ∫dk(2π)3∑nf(En,k)Ωγ(n,k)(1)\sigma_{\alpha\beta}=-\frac{e^2}{\hbar}\epsilon_{\alpha\beta\gamma}\int\frac{d k}{(2\pi)^3}\sum_n f(E_{n,\boldsymbol{k}})\Omega_\gamma(n,\boldsymbol{k}) \tag*{(1)}

where ϵαβγ\epsilon_{\alpha\beta\gamma} is Levi-Civita symbol. En,kE_{n,\boldsymbol{k}} and Ωγ\Omega_\gamma are the energy eigenvalue and Berry curvature, respectively, of the band with quantum number nn. f(En,k)f(E_{n,\boldsymbol{k}}) is the Fermi-Dirac distribution function. At zero temperature, the summation in Eq. 1 reduces to a sum over occupied bands. To simulate AHC within DFT, we constructed a tight-binding Hamiltonian from maximally localized Wannier functions, which were obtained through Wannier90 [38]. Wannier interpolation of Berry curvature and AHC were postprocessed using the “pruned FFT”-based WannierBerri [39] package. Note that rapid variations of Berry curvature are taken into account by integrating the Brillouin Zone on a dense kk-mesh of 125×125×1125\times125\times1 and with recursive adaptive mesh refinement. We conducted symmetry analysis using Ref. [40], FINDSYM [41], Bilbao Crystallographic Server [42], MAGNDATA [43], and AMCHECK [44].

Results and Discussion

Symmetry analysis for AHE in 2D AMs

Since AMs display spin-splitting of electronic structure without spin-orbit coupling (SOC), they are best characterized by spin group symmetries [R1∥R2][R_1\Vert R_2], where R1R_1 and R2R_2 symmetry operations act on the decoupled spin and real space, respectively [3] [45, 46, 47]. Just like in the case of AFMs, symmetry dictates a net zero magnetization in AMs. However, in AFMs, the spin-up and -down bands are degenerate. It is instructive to see which symmetries (in terms of spin group formalism) would result in a traditional AFM versus an AM. Collinear magnets, including AMs, always have spin-only symmetry [C2∥T][C_2\Vert T], where C2C_2 is a two-fold rotation around the axis perpendicular to the collinear spins, followed by inversion in spin-space [45]. The [C‾2∥T][\overline{C}_2\Vert T] transforms energy eigenstates

as [C‾2∥T]E(s,k)=E(s,−k)[\overline{C}_2\Vert T]E(s,\boldsymbol{k})=E(s,-\boldsymbol{k}), leading to even-parity spin splitting. The [C2∥T][C2∥P][C_2\Vert T][C_2\Vert P] and [C2∥τ][C_2\Vert\tau] transform the energy eigenstates from E(s,k)E(s,\boldsymbol{k}) to E(−s,k)E(-s,\boldsymbol{k}), which lead to spin degeneracy at an arbitrary kk-point. The C2C_2 operation is the two-fold rotation about an axis perpendicular to collinear spins in spin space and, for simplicity, can be interpreted as spin space inversion. Additionally, for the 2D case (kz=0k_z=0), the spin-group symmetries [C2∥Mz][C_2\Vert M_z] and [C2∥C2z][C_2\Vert C_2^z] also lead to spin degeneracy (see Sec. I of Supplemental Material (SM) [48]). Overall, to obtain altermagnetism in 2D materials, opposite spin sublattices should not be connected by crystallographic PP, τ\tau, C2zC_2^z, and MzM_z. However, there should exist at least one crystallographic symmetry connecting opposite spin sublattices to have symmetry-enforced net zero magnetization. The possible symmetry options are in-plane two-fold rotation (i.e., [C2∥C2z][C_2\Vert C_2^z]), vertical mirror plane (i.e., [C2∥Mx][C_2\Vert M_x]), and out-of-plane four-fold rotation ([C2∥C4z][C_2\Vert C_4^z]). The 2D AM case with [C2∥Mx][C_2\Vert M_x] symmetry is highlighted in Figure 1.

Schematic diagram showing how symmetries determine the presence or absence of AHE in 2D altermagnets

Figure 1. Schematic diagram showing how symmetries determine the presence or absence of AHE in 2D altermagnets. In this example, the spin-group symmetry [C2∥Mx][C_2\Vert M_x] connects the opposite spin-sublattices. (a) MxM_x symmetry is satisfied when the Néel vector (N\boldsymbol{N}) is perpendicular to the direction of the mirror plane. Note that the direction of mirror plane refers to the direction normal to the mirror (here, xx-direction). (b) MxM_x symmetry forbids the AHE (see Table I). (c) N\boldsymbol{N} being parallel to the xx-direction violates the MxM_x symmetry and (d) thus the AHE is symmetry allowed. Since N\boldsymbol{N}, JH\boldsymbol{J}_{\mathrm{H}}, and E\boldsymbol{E} all lie in the same plane (here, the xx-yy plane), this is referred to as the in-plane AHE.

The symmetry analysis for the relativistic AHE requires the considerations of magnetic symmetry operations acting on coupled spin and real spaces. From the symmetry perspective, the Hall vector, σH\boldsymbol{\sigma}_{\mathrm{H}}, transforms like a pseudovector, similar to magnetization. Therefore, the magnetic symmetry operations impose certain constraints on allowed components, σij\sigma_{ij}, of the system (see Table I and Sec. II of SM [48]). In particular, all components of the AHC tensor are symmetry forbidden if a material (2D or 3D) possesses either of the TT, PTPT, TC3,4,6x,zTC_{3,4,6}^{x,z} or TS4,6x,zTS_{4,6}^{x,z} symmetries. Further constraints determining non-zero AHC are imposed by the 2D nature of the 2D AMs. As the Hall current is restricted to the plane of the 2D materials, the only component that is experimentally relevant is σxy\sigma_{xy} (taking zz as the out-of-plane axis). Hence, the presence of CnxC_n^x, MxM_x, TMzTM_z, and TC2zTC_2^z symmetries suppresses AHE in 2D materials as these symmetries result in zero σxy\sigma_{xy} [see Table I]. Overall, our symmetry analysis shows that three of the following conditions must be simultaneously satisfied to obtain a 2D AM with AHE: (i) absence of [C2∥P][C_2\parallel P], [C2∥τ][C_2\parallel\tau], [C2∥Mz][C_2\parallel M_z], and [C2∥C4z][C_2\parallel C_4^z] spin group symmetries that will otherwise indicate that we have a conventional AFM, (ii) presence of at least one of [C2∥C2z][C_2\parallel C_2^z], [C2∥Mx][C_2\parallel M_x], and [C2∥C4z][C_2\parallel C_4^z] spin group symmetries, ensuring altermagnetism, and (iii) absence of TT, PTPT, TC3,4,6x,zTC_{3,4,6}^{x,z}, TS4,6x,zTS_{4,6}^{x,z}, CnxC_n^x, MxM_x, TMzTM_z, and TC2zTC_2^z magnetic symmetries, which ensures non-zero σxy\sigma_{xy}. These conditions for obtaining 2D AMs with AHE are partially related. For example, [C2∥P][C_2\parallel P] and PTPT occur simultaneously in collinear magnets. These conditions make observing AHE difficult in 2D AMs, and have largely remained unexplored.

The magnetic symmetries depend on the Néel vector orientation with respect to the crystal symmetry. The Néel vector serves as an order parameter for AFMs and offers a robust nonvolatile approach to modifying magnetic symmetries [49]. Since, AMs also have net zero magnetization, the Néel vector could serve as a natural way to modify AHE. For example, consider the case of a 2D AM with [C2∥Mx][C_2\parallel M_x] symmetry, where opposite spin sublattices are connected through a vertical mirror plane (see Figure 1). When the Néel vector is perpendicular to the xx-direction, the magnetic point group of the AM will contain MxM_x, leading to vanishing σxy\sigma_{xy} [Figure 1(a) and Figure 1(b)]. However, aligning the Néel vector along the xx-direction will lead to the breaking of the MxM_x symmetry [Figure 1(c)]. Hence, the presence of the Néel vector along the xx-direction will allow σxy\sigma_{xy} [Figure 1(d)]. It is worth mentioning that MxM_x and C2xC_2^x impose the same condition on σxy\sigma_{xy} [see Table I]. Therefore, the presence of [C2∥C2x][C_2\parallel C_2^x] will allow (forbid) σxy\sigma_{xy} when the Néel vector is parallel (perpendicular) to xx direction. Following the same approach, we have classified the 2D AMs based on whether the AHE is allowed for the different Néel vector orientations, spin-momentum coupling and their nontrivial spin layer group (SLG) (see Table II and Sec. III of SM [48]). Interestingly, dd-wave AMs with nontrivial SLG 22/2m2^2/2m and ii-wave AMs with nontrivial SLG 13ˉ2m1\bar{3}2m allow for an in-plane AHE, where N\mathbf{N}, JHJ_{\mathrm{H}}, and E\mathbf{E} lie in the x−yx-y plane. Additionally, AMs with nontrivial SLGs 2m2m1m2m^2m^1m, 24/1m2^4/{}^1m, 24/1m2m2m2^4/{}^1m^2m^2m, and 21/1m2m2m2^1/{}^1m^2m^2m do not show an in-plane AHE due to the presence of TC2zTC_2^z symmetry. In accordance to symmetry analysis, σxy\sigma_{xy} is always forbidden for 2D AMs when Néel vector points along the zz-direction.

Nontrivial SLGSpin-momentum couplingNontrivial spin group operationsSymmetry-allowed σxy\sigma_{xy} termsσxy\sigma_{xy} N⊥zN \perp zσxy\sigma_{xy} N∥zN \parallel zExamples
22/2m{}^{2}2/{}^{2}mdd-wave[C2∥C2x][C_{2}\Vert C_{2}^{x}]Nx1,3N_{x}^{1,3}, NxNyNzbN_{x}N_{y}N_{z}^{b} (a+b=2a+b=2)✓✗ (C2xC_{2}^{x})RuF4_4[23], dd-MnPSe3_3 [⋆]
2m2m1m{}^{2}m{}^{2}m{}^{1}mdd-wave[C2∥C2x][C_{2}\Vert C_{2}^{x}], [E∥C2z][E\Vert C_{2}^{z}]NxNyNzN_{x}N_{y}N_{z}✗ (TC2zT C_{2}^{z})✗ (C2xC_{2}^{x})MnTeMoO6_6 [19]
24/1m{}^{2}4/{}^{1}mdd-wave[C2∥C4z][C_{2}\Vert C_{4}^{z}], [E∥C2z][E\Vert C_{2}^{z}]Nx2Nz−Ny2NzN_{x}^{2}N_{z}-N_{y}^{2}N_{z}, NxNyNzN_{x}N_{y}N_{z}✗ (TC2zT C_{2}^{z})✗ (C2xC_{2}^{x})
24/1m2m1m{}^{2}4/{}^{1}m{}^{2}m{}^{1}mdd-wave[C2∥C4z][C_{2}\Vert C_{4}^{z}], [E∥C2z,C2x][E\Vert C_{2}^{z},C_{2}^{x}]Nx2Nz−Ny2NzN_{x}^{2}N_{z}-N_{y}^{2}N_{z}✗ (TC2zT C_{2}^{z})✗ (C2xC_{2}^{x})VSe2_2O [50], CrO [51]
14/1m2m2m{}^{1}4/{}^{1}m{}^{2}m{}^{2}mgg-wave[C2∥C2x][C_{2}\Vert C_{2}^{x}], [E∥C4z,C2z][E\Vert C_{4}^{z},C_{2}^{z}]—✗ (TC2zT C_{2}^{z})✗ (C2xC_{2}^{x})VP2_2H8_8(NO4_4)2_2 [19]
132m{}^{1}3{}^{2}mii-wave[C2∥C2x][C_{2}\Vert C_{2}^{x}], [E∥C3z][E\Vert C_{3}^{z}]Nx3−3NxNy2N_{x}^{3}-3N_{x}N_{y}^{2}✓✗ (C2xC_{2}^{x})ii-MnPSe3_3 [⋆]
16/1m2m2m{}^{1}6/{}^{1}m{}^{2}m{}^{2}mii-wave[C2∥C2x][C_{2}\Vert C_{2}^{x}], [E∥C3z,C2z][E\Vert C_{3}^{z},C_{2}^{z}]—✗ (TC2zT C_{2}^{z})✗ (C2xC_{2}^{x})

Table II. AHE in 2D AMs with different nontrivial spin-Layer group (SLG) symmetry. The superscripts 1 and 2 denote symmetry operations connecting atoms with the same and opposite spin magnetizations, respectively. Spin-momentum coupling is categorized based on the number of spin-degenerate nodal lines in the band dispersion. Specifically, dd-, gg-, and ii-wave AMs exhibit 2, 4, and 6 spin-degenerate nodal lines passing through the Γ\Gamma point, respectively, in the (kx−ky)(k_x-k_y) momentum space. The nontrivial spin group symmetry operations are highlighted for each case. Note that the spin group operations [E∥E][E\Vert E] and [E∥P][E\Vert P] are excluded as they impose no restrictions on the AHC. Symmetry-invariant terms describe the functional dependence of the AHC σxy\sigma_{xy} on the Néel vector components (Nx,Ny,Nz)(N_x,N_y,N_z). The table also specifies whether σxy\sigma_{xy} is allowed when the Néel vector is perpendicular or parallel to the out-of-plane direction (zz). For forbidden cases, the relevant magnetic symmetry responsible is also indicated. Symmetry invariants up to cubic in NiN_i are included. Note that symmetry invariants may encompass higher-order terms, though their contributions are expected to be weak. Additionally, examples of well-known 2D AMs belonging to different SLGs are also tabulated. Entries marked with [⋆] denote examples from this work.

Bilayer AMs as prototypical candidates

We exemplify the symmetry predictions of AHE in the 2D AMs using DFT simulations. For this, we chose the prototypical experimentally-synthesized AFM materials – MnPSe3_3 [52–54] and MnSe [33, 55–57] monolayers. In what follows, we provide detailed results for MnPSe3_3, while details of our calculations for MnSe are included in the supplemental material [48]. Both MnPSe3_3 and MnSe monolayers form a large class of 2D materials with G and A type AFM order, respectively, and a Néel transition temperature of ∼\sim75 K [55, 58]. The PTPT symmetry in MnPSe3_3 and MnSe monolayers enforces spin-degeneracy and forbids the AHE effect (see Sec. IV of SM [48]). To obtain AM, we break PTPT symmetry of the monolayer MnPSe3_3 and MnSe by using the bilayer stacking approach [see Fig. 2(a)] [18]. We created different bilayer stacking by first taking the top layer to be the mirror reflection of the lower layer, followed by the translation of the upper layer in the basal plane. We have also used two different magnetic configurations with intralayer AFM, namely, M↑↓↑↓ and M↑↓↓↑ (the four arrows denote the magnetization directions of Mn atoms, with first two arrows being used for the lower layer and the last two arrows for the upper layer). The configurations M↑↓↑↓ and M↑↓↓↑ are almost degenerate (differing by around 0.22 meV), signifying the weak interlayer exchange interaction. Further, DFT calculations were performed for potential energy surfaces of various high symmetry stackings [see Fig. 2(b)] for MnPSe3_3. There are 6 degenerate lowest-energy configurations. These stackings are equivalent by symmetry and are obtained through translation of the upper layer by 13a\frac{1}{3}a, 13b\frac{1}{3}b, 23a\frac{2}{3}a, 23b\frac{2}{3}b, 13a+13b\frac{1}{3}a+\frac{1}{3}b, and 23a+23b\frac{2}{3}a+\frac{2}{3}b. Similarly, three high-energy stacking configurations, obtained through translation of the upper layer by 00, 13a+23b\frac{1}{3}a+\frac{2}{3}b, and 23a+13b\frac{2}{3}a+\frac{1}{3}b, are degenerate. Therefore, we take representative cases of bilayer MnPSe3_3 obtained through 13a+13b\frac{1}{3}a+\frac{1}{3}b [see Fig. 2(c)] and 13a+23b\frac{1}{3}a+\frac{2}{3}b [see Fig. 2(d)] and where these stackings as dd-MnPSe3_3 and ii-MnPSe3_3, respectively. The rationale behind this unconventional nomenclature is elaborated in the following paragraph. The most stable Néel ordering is M↑↓↑↓ and M↑↓↓↑ for dd-MnPSe3_3 and ii-MnPSe3_3, respectively, and therefore used in our simulations. Although, ii-MnPSe3_3 is not the most stable bilayer stacking [see Fig. 2(d)], it serves a qualitative analysis of AHE in the materials sharing structural similarity. For instance, twisted bilayers (tb) of hexagonal materials (such as tb-NiCl2_2 [21], tb-MnBi2_2Te4_4 [24], and tb-MnPSe3_3 [25]) have the same nonmagnetic and magnetic point group as of ii-MnPSe3_3. However, performing DFT+U+SOC simulations on twisted bilayers are computationally formidable task owing to supercell size.

We computed the electronic bands of dd-MnPSe3_3 and ii-MnPSe3_3 without SOC [see Figs. 3(a) and 3(b)]. The spin degeneracy in the energy bands can be explained through spin group symmetry operations [3]. The presence of [C2∥O][C_2\Vert O] symmetry leads to [C2∥O]E(s,k)=E(−s,O−1k)[C_2\Vert O]E(s,\mathbf{k})=E(-s,O^{-1}\mathbf{k}). Therefore, the bands are spin degenerate along the paths for which Ok=kO\mathbf{k}=\mathbf{k} or Ok=k+GO\mathbf{k}=\mathbf{k}+\mathbf{G}, where G\mathbf{G} is a reciprocal lattice vector. For dd-MnPSe3_3, the spin-up and spin-down states are degenerate for the bands along the directions perpendicular and parallel to the mirror plane due to [C2∥M−3x+y][C_2\Vert M_{-\sqrt{3}x+y}] and [C2‾∥T][C2∥M−3x+y][\overline{C_2}\Vert T][C_2\Vert M_{-\sqrt{3}x+y}], respectively (see Sec. V of SM [48] for constant energy contours). AMs with two spin degenerate nodal lines are classified as dd-wave AMs, and dd-MnPSe3_3 belong to that class [3]. Similarly, the energy bands in ii-MnPSe3_3 are spin degenerate along all possible high symmetry directions Γ−M\Gamma-M, Γ−K\Gamma-K, and M−KM-K due to presence of three [C2∥C2′][C_2\Vert C_2']-type spin group symmetries (Figure 2(d)) and their combinations with spin-only symmetry [C2‾∥T][\overline{C_2}\Vert T]. This results in the ii-wave altermagnetism in the ii-MnPSe3_3. Note that although bands are nondegenerate at the general k\mathbf{k}-point, the sum of spin splittings throughout the BZ, ∑BZE(s,k)−E(−s,k)\sum_{\mathrm{BZ}} E(s,\mathbf{k})-E(-s,\mathbf{k}), is zero for each case (see Figure 3(a) and Figure 3(b)).

Generation of AM through bilayer stackings

Figure 2. Generation of AM through bilayer stackings. (a) The upper layer is obtained by taking the horizontal mirror reflection of the lower layer followed by in-plane translation. The bilayer stacking may have [C2∥My][C_2\parallel M_y] and [C2∥C2y][C_2\parallel C_2^y]-like spin group symmetries depending on the shifting vector and the constituent monolayer. (b) The energy distribution of the bilayer MnPSe3_3 as a function of the shifting of the upper layer. (c) and (d) highlights the two high-symmetry stackings obtained by shifting the upper layer by 13a+13b\frac{1}{3}\mathbf{a}+\frac{1}{3}\mathbf{b} (d-MnPSe3_3) and 13a+23b\frac{1}{3}\mathbf{a}+\frac{2}{3}\mathbf{b} (i-MnPSe3_3), respectively, with representative Néel vector along xx-direction. The blue and yellow spheres denote the Mn atoms from the top and bottom layers, respectively. We have omitted the P and Se atoms for clear illustration (see Figure S2 in the SM [48] for complete structure). The position of structures in (c) and (d) are also highlighted in potential energy surfaces in (b) with up- and down-triangle, respectively. The spin group symmetry operations are also indicated for each case.

Valence bands and anomalous Hall conductivity of d-MnPSe3 and i-MnPSe3

Figure 3. Valence bands of (a) dd-MnPSe3_3 and (b) ii-MnPSe3_3 bilayer without inclusion of SOC. The red dashes and blue dots represent spin-up and spin-down bands, respectively. Anomalous Hall conductivity σxy\sigma_{xy} of (c) dd-MnPSe3_3 and (d) ii-MnPSe3_3 as a function of Fermi energy for Néel vector pointing along different directions. The zero Fermi energy corresponds to valence band maximum.

The AHC and Néel vector relationship

We calculate the AHC, σxy\sigma_{xy}, as a function of the Fermi energy for dd-MnPSe3_3 and ii-MnPSe3_3 for different Néel vector, N\mathbf{N}, orientations (Figure 3(c) and Figure 3(d)). The σxy\sigma_{xy} shows a strong dependence on the N\mathbf{N} orientation and is absent when N\mathbf{N} is along the zz-direction. The anomalous Hall response is strongest for N\mathbf{N} along the xx-direction. The σxy\sigma_{xy} for dd-MnPSe3_3 develops a strong peak of 202 S/cm around −0.23-0.23 eV, while σxy\sigma_{xy} of ii-MnPSe3_3 develops a slightly weaker peak of 25 S/cm around −0.17-0.17 eV. Note that we observe a very sharp peak for ii-MnPSe3_3 around −0.25-0.25 eV and it may be difficult to achieve experimentally as it requires precise control of the Fermi energy. Interestingly, σxy\sigma_{xy} is forbidden for N\mathbf{N} aligned along the yy-direction for ii-MnPSe3_3, while it shows an intermediate response for dd-MnPSe3_3. To understand this dependence of σxy\sigma_{xy} on the Néel vector, we plot the anomalous Hall peaks for different N\mathbf{N} orientations in the xx-yy and xx-zz planes (see Figure 4). For the xx-yy plane, the AHE is absent in dd-MnPSe3_3 for the Néel vector along 60∘/240∘60^\circ/240^\circ due to the presence of the vertical mirror M−3x+yM_{-\sqrt{3}x+y} (Figure 4(a)). Similarly, the AHE is absent for ii-MnPSe3_3 when N\mathbf{N} is perpendicular to any of the two-fold rotation symmetries (30∘30^\circ, 90∘90^\circ, 150∘150^\circ, 210∘210^\circ, 270∘270^\circ, and 330∘330^\circ) (Figure 4(c)). When N\mathbf{N} points in the xx-direction, the AHE is forbidden due to presence of the M−3x+yM_{-\sqrt{3}x+y} and C2xC_2^x in dd-MnPSe3_3 and ii-MnPSe3_3, respectively (see Figure 4(b) and Figure 4(d)).

Anomalous Hall conductivity of d-MnPSe3 and i-MnPSe3 as a function of Néel vector orientation

Figure 4. The anomalous Hall conductivity (σxy\sigma_{xy}) of dd-MnPSe3_3 with dd-wave AM as a function of the Néel vector orientation in (a) xx-yy plane and (b) xx-zz plane at −0.23-0.23 eV. Similarly, (c) and (d) show the AHC for ii-MnPSe3_3 with ii-wave AM at −0.17-0.17 eV. The regions with positive and negative σxy\sigma_{xy} are highlighted with light red and light blue. The green dots represent the calculated DFT values while black curves are the fits using models in Eq. 2.

Furthermore, we understand the periodicity by writing the general functional form of σxy\sigma_{xy} as σxy(N)=∑m,n,rλmrnNxmNynNzr\sigma_{xy}(\mathbf{N})=\sum_{m,n,r}\lambda_{mrn}N_x^mN_y^nN_z^r, where mm, nn, and rr are whole numbers. Some components of λmnr\lambda_{mnr} are forbidden due to symmetries. For instance, m+n+rm+n+r is never even as it would violate the Onsager reciprocity relation: σxy(N)=σyx(−N)\sigma_{xy}(\mathbf{N})=\sigma_{yx}(-\mathbf{N}) [59]. We use the “method of invariants” [60] [σxy(ON)=σxy(N)][\sigma_{xy}(O\mathbf{N})=\sigma_{xy}(\mathbf{N})] to obtain symmetry allowed components λmnr\lambda_{mnr}. For this, we treat N\mathbf{N} as an extrinsic parameter instead as an intrinsic one [61, 62]. Recall that N\mathbf{N} transforms like N′=±Det⁡(O)D(O)N\mathbf{N}'=\pm\operatorname{Det}(O)D(O)\mathbf{N}, where D(O)D(O) and Det⁡(O)\operatorname{Det}(O) are the matrix representation and determinant of OO. The ++ (−-) sign is taken if OO exchanges the same (opposite) spin sublattices (see Figure 1(a) and Figure 1(b)). Table II summarizes the symmetry allowed profiles of σxy(N)\sigma_{xy}(\mathbf{N}) for 2D AMs with different spin group symmetry operations and nontrivial SLG (see also Sec. III of SM [48]). dd-MnPSe3_3 and ii-MnPSe3_3 have nontrivial SLG 22/2m2^2/2m and 13‾2m1\overline{3}2m, respectively, and the σxy−N\sigma_{xy}-\mathbf{N} relationship can be expressed as

σxy22/2m=λ100Nx+λ300Nx3+λ111NxNyNz+λ120NxNy2+λ102NxNz2(2)\sigma_{xy}^{2^2/2m}=\lambda_{100}N_x+\lambda_{300}N_x^3+\lambda_{111}N_xN_yN_z+\lambda_{120}N_xN_y^2\\ +\lambda_{102}N_xN_z^2 \tag*{(2)}
σxy13‾2m=λ300(Nx3−3NxNy2)+λ302(Nx3−3NxNy2)Nz2.(3)\sigma_{xy}^{1\overline{3}2m}=\lambda_{300}(N_x^3-3N_xN_y^2)+\lambda_{302}(N_x^3-3N_xN_y^2)N_z^2. \tag*{(3)}

In Figure 4, we fit the calculated DFT values of σxy\sigma_{xy} with the models in Eq. 2. Note that for dd-MnPSe3_3 the mirror plane is M−3x+yM_{-\sqrt{3}x+y} and not MxM_x. Therefore, we have rotated the models accordingly while fitting. The models provide good agreement with the DFT results. However, the fit shows slight deviations from the DFT results, which may be due to the structural distortions relative to the ideal symmetric configurations, and/or because we have ignored the higher-order terms in Eq. 2. Note that the σxy\sigma_{xy} of dd-MnPSe3_3 depends on linear and cubic terms in the Néel vector NN [Fig. 4(a)], while ii-MnPSe3_3 exhibits a purely cubic dependence on NN [Fig. 4(c)]. Consequently, the strength of AHC in ii-MnPSe3_3 is smaller than that in dd-MnPSe3_3, with three AHC-forbidden directions in the xx-yy plane of the Néel vector space.

The dependence of σxy\sigma_{xy} on NN is derived for all 2D altermagnetic nontrivial SLGs in Table II. The AHE is forbidden for all gg-wave 2D AMs due to the presence of [E∥C4z][E\Vert C_4^z] symmetry, while the [E∥C2z,C2z][E\Vert C_2^z,C_2^z] symmetries suppress the AHE in ii-wave 2D AMs with the nontrivial SLG 16/1ˉm2m2m16/\bar{1}m^2m^2m. The presence of [E∥C2z][E\Vert C_2^z], [E∥C4z][E\Vert C_4^z], or [C2∥∥C2z][C_{2\parallel}\Vert C_2^z] leads to the presence of magnetic TC2zTC_{2z} symmetry, which forbids the in-plane AHE. Most commonly, 2D magnets have an easy axis of magnetization that lies completely in-plane [34, 57] or out-of-plane [63, 64], and in such a scenario, the AHE can be observed for only 22/m2^{2}/m and 13ˉ2m1\bar{3}^{2}m out of the seven nontrivial SLGs. Interestingly, the dependence of σxy\sigma_{xy} on the Néel vector space is similar to that of the Berry curvature multipoles on kk-space [65] and the spin magnetization multipoles on real space [66].

The quantum origin of the large AHC lies in the SOC-induced avoided crossings, which act as opposite poles of Berry curvature (see § 11 of SM [48]). The highest σxy\sigma_{xy} is achieved if the Fermi energy lies in the middle of the poles. In the case of the antiferromagnetic MnSe bilayer, with the translation of one layer relative to the other by 13a+13b\frac{1}{3}a+\frac{1}{3}b, we obtain effects similar to those in the dd-MnPSe3_3 bilayer, including altermagnetism (see § 12 of SM [48]). Note that bilayer stacking of AFMs may also lead to weak ferromagnetism [67, 68] where the opposite spin sublattices are not connected by any symmetry. For instance, the polar stacking of bilayer MnSe with basal plane translated by 13a+23b\frac{1}{3}a+\frac{2}{3}b leads to weak ferromagnetism in the structure. In such cases, the anomalous Hall response is FM-like, where the AHE is also observed in the plane perpendicular to Néel vector. The weakly FM bilayer MnSe shows σxy\sigma_{xy} when the Néel vector points along the zz-direction similar to conventional FMs [69, 70] (see § 13 of SM [48]) and differ from the AHE in 2D AMs.

SUMMARY

We have established the principles for obtaining the AHE in 2D AMs. Our analysis shows that TC2zTC_{2z} and C2xzC_{2x}^{z} are the most common magnetic point group symmetries that suppress the AHE in 2D AMs. Although the AHE is a relativistic effect, spin group symmetries are effective in explaining the unconventional periodicity in the in-plane anomalous Hall response, which can be used to detect the altermagnetic order and Néel vector reversal. We also show that the AHE is forbidden for gg-wave 2D AMs. Our symmetry predictions are supported by first-principles DFT simulations for bilayer MnPSe3_3 with two-different layer-stacking geometries used as prototypical candidates for the dd- and ii-wave AMs. Overall, 2D AMs are promising for the miniaturization of spintronic memory devices, with the Néel vector serving as the write-in mechanism and in-plane AHE for read-off of the information. For the purpose of application in spintronics, 2D AMs that have an in-plane easy axis of magnetization will form a suitable choice. The spin-momentum coupling, and hence, the altermagnetic order can be controlled in twisted magnetic bilayers, depending on the symmetry of the constituent monolayers [24], providing an exceptional platform to achieve the AHE in experiments. Additionally, external parameters such as electric fields or strain may induce AHE in otherwise forbidden AMs [71]. The approach provided here can be extended to other similar transverse effects, such as the anomalous Nernst effect [72] and the nonlinear Hall effect [73]. We anticipate that the theoretical findings of this work will enrich the field of altermagnetic spintronics [4].

ACKNOWLEDGMENTS

This material is based upon work supported by the Air Force Office of Scientific Research under award number FA9550-23-1-0679 and the National Science Foundation Grant No. DMR-1752840. This work used the Expanse cluster at SDSC through allocation PHY180014 from the Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program, which is supported by National Science Foundation Grants Nos. 2138259, 2138286, 2138307, 2137603, and 2138296, and Maryland Advanced Research Computing Center.

Supplemental Material for “Spontaneous Anomalous Hall Effect in Two-Dimensional Altermagnets”

Sajjan Sheoran; Pratibha Dev

2D Altermagnetism: group theoretic considerations

The main text describes symmetry considerations that lead to the emergence of 2D altermagnetism. In this subsection, we provide greater details of how group theoretic considerations were used to obtain our results. If the Hamiltonian of a system remains invariant under a certain symmetry (OO), it transforms the energy eigenstate (EE) as OE(s,k)=E(s′,k′)OE(s,\mathbf{k}) = E(s',\mathbf{k}'), where s′=Oss' = Os and k′=Ok\mathbf{k}' = O\mathbf{k}. In the nonrelativistic case, spin and real space are decoupled. In this case, the symmetry transformation, OO, can be written as O=[O1∥O2]O = [O_1 \mathbin\Vert O_2], where O1O_1 and O2O_2 act only on the spin and real spaces, respectively. Generally, [O1∥O2][O_1 \mathbin\Vert O_2] can take one of four forms: [E∥E][E \mathbin\Vert E], [O1∥E][O_1 \mathbin\Vert E], [E∥O2][E \mathbin\Vert O_2], and [O1∥O2][O_1 \mathbin\Vert O_2]. Here, [E∥E][E \mathbin\Vert E] represents the trivial identity in both spaces, [O1∥E][O_1 \mathbin\Vert E] corresponds to spin-only symmetry, [E∥O2][E \mathbin\Vert O_2] represents crystallographic symmetry (nonmagnetic point group), and [O1∥O2][O_1 \mathbin\Vert O_2] represents nontrivial spin group symmetry. Collinear magnets can exhibit spin-only symmetries, such as C∞C_{\infty} (arbitrary rotation around the collinear spin axis) and Cˉ2\bar{C}_2 (a two-fold rotation around an axis perpendicular to the collinear spin axis, followed by inversion) in addition to the trivial identity. The symmetry C∞C_{\infty} allows the separation of spin-up and spin-down bands. Meanwhile, spin-space inversion in Cˉ2\bar{C}_2 occurs with the time-reversal operation, which also acts on the real space. As a result, collinear magnets always exhibit [Cˉ2∥T][\bar{C}_2 \mathbin\Vert T] symmetry. For collinear magnets without spin-orbit coupling (SOC), the general symmetry operation can be written as:

([E∥E]+[C∞∥E]+[Cˉ2∥T]+[C∞∥E][Cˉ2∥T])[O1∥O2],([E \mathbin\Vert E] + [C_{\infty} \mathbin\Vert E] + [\bar{C}_2 \mathbin\Vert T] + [C_{\infty} \mathbin\Vert E][\bar{C}_2 \mathbin\Vert T])[O_1 \mathbin\Vert O_2],

where the terms inside the first set of brackets represent the spin-only group, and [O1∥O2][O_1 \mathbin\Vert O_2] is a nontrivial spin group symmetry dependent on the crystal and magnetic structure. For example, if opposite spin sublattices are connected by inversion (PP) or translation (τ\tau), the nontrivial spin group symmetry is [Cˉ2∥P][\bar{C}_2 \mathbin\Vert P] or [Cˉ2∥τ][\bar{C}_2 \mathbin\Vert\tau], where C2C_2 represents a two-fold rotation about an axis perpendicular to the collinear spin axis. The C2C_2 acts on spin-space same as spin-space inversion (the difference being that it is not accompanied by time-reversal (TT)). The [Cˉ2∥T][Cˉ2∥P][\bar{C}_2 \mathbin\Vert T][\bar{C}_2 \mathbin\Vert P] and [Cˉ2∥τ][\bar{C}_2 \mathbin\Vert\tau] symmetries transform energy eigenstates as follows:

[Cˉ2∥T][Cˉ2∥P]E(s,k)=[Cˉ2∥T]E(−s,−k)=E(−s,k)(4)[\bar{C}_2 \mathbin\Vert T][\bar{C}_2 \mathbin\Vert P]E(s,\mathbf{k}) = [\bar{C}_2 \mathbin\Vert T]E(-s,-\mathbf{k}) = E(-s,\mathbf{k}) \tag*{(4)}
[Cˉ2∥τ]E(s,k)=E(−s,k)(5)[\bar{C}_2 \mathbin\Vert\tau]E(s,\mathbf{k}) = E(-s,\mathbf{k}) \tag*{(5)}

Therefore, [Cˉ2∥P][\bar{C}_2 \mathbin\Vert P] and [Cˉ2∥τ][\bar{C}_2 \mathbin\Vert\tau] lead to spin degeneracy at arbitrary an k\mathbf{k}-point. For the 2D materials, k=kxi+kyj\mathbf{k} = k_x\mathbf{i} + k_y\mathbf{j} (kz=0k_z = 0). The transformations of [Cˉ2∥Mz][\bar{C}_2 \mathbin\Vert M_z] and [Cˉ2∥T][Cˉ2∥C2z][\bar{C}_2 \mathbin\Vert T][\bar{C}_2 \mathbin\Vert C_{2z}] act on energy eigenstates as follows:

[Cˉ2∥Mz]E(s,k)=E(−s,k)(6)[\bar{C}_2 \mathbin\Vert M_z]E(s,\mathbf{k}) = E(-s,\mathbf{k}) \tag*{(6)}
[Cˉ2∥T][Cˉ2∥C2z]E(s,k)=[Cˉ2∥T]E(−s,−k)=E(−s,k)(7)[\bar{C}_2 \mathbin\Vert T][\bar{C}_2 \mathbin\Vert C_{2z}]E(s,\mathbf{k}) = [\bar{C}_2 \mathbin\Vert T]E(-s,-\mathbf{k}) = E(-s,\mathbf{k}) \tag*{(7)}

Here, the operator MzM_z represents mirror symmetry about the z=0z=0 plane, while C2zC_{2z} denotes a two-fold rotation about the zz-axis in real space. Consequently, the symmetries [C2∥C2z][C_2\Vert C_{2z}] and [C2∥Mz][C_2\Vert M_z] enforce spin degeneracy in 2D materials. To break spin degeneracy in 2D materials, opposite spin sublattices must not be related by PP, τ\tau, MzM_z, or C2zC_{2z}.

For 2D altermagnetism to emerge, there must be specific symmetries that connect opposite spin sublattices. While, in principle, any symmetry except PP, τ\tau, MzM_z, and C2zC_{2z} can connect opposite spin sublattices, not all such symmetries are practical in 2D materials. For example, 2D materials cannot exhibit C3xC_3^x symmetry. Similarly, although C3zC_3^z symmetry is allowed in 2D materials, it cannot connect opposite spin sublattices, as it would lead to [C2∥C3z][C2∥C3z][C2∥C3z]=[C2∥E][C_2\Vert C_3^z][C_2\Vert C_3^z][C_2\Vert C_3^z]=[C_2\Vert E], which means opposite spin sublattices are connected by identity (not possible for magnets). Similarly it can be shown that the only symmetries that can connect opposite spin sublattices in 2D materials are C2xC_2^x, C4zC_4^z, and MxM_x. In spin group notation, the corresponding symmetries are [C2∥C2x][C_2\Vert C_2^x], [C2∥C4z][C_2\Vert C_4^z], and [C2∥Mx][C_2\Vert M_x]. It is worth mentioning that symmetry [C2∥S4z][C_2\Vert S_4^z] may also connect opposite spin sublattices, however, it is redundant as it can be written as combination of [E∥P][E\Vert P] and [C2∥C4z][C_2\Vert C_4^z] and we are concerned only about generators.

Symmetry constraints on AHE

In what follows, we provide our procedure for determining the constraints on the AHC tensor components, which are imposed by unitary and antiunitary symmetry operations (see Table I of the main text). The AHC is expressed as:

σαβ=−ϵαβγe2ℏ∫dk(2π)3∑nf(En,k)Ωγ(n,k),(8)\sigma_{\alpha\beta}=-\epsilon_{\alpha\beta\gamma}\frac{e^2}{\hbar}\int\frac{d\mathbf{k}}{(2\pi)^3}\sum_n f(\mathcal{E}_{n,\mathbf{k}})\Omega_\gamma(n,\mathbf{k}), \tag*{(8)}

where Ωγ(n,k)\Omega_\gamma(n,\mathbf{k}) is the Berry curvature of band nn at the k\mathbf{k}-point, En,k\mathcal{E}_{n,\mathbf{k}} is the energy eigenvalue, ϵαβγ\epsilon_{\alpha\beta\gamma} is the Levi-Civita symbol, and f(En,k)f(\mathcal{E}_{n,\mathbf{k}}) is the Fermi-Dirac distribution function. The Berry curvature and σH=(σzy,σxz,σyx)\boldsymbol{\sigma}_H=(\sigma_{zy},\sigma_{xz},\sigma_{yx}) transform as pseudovectors. Also, the AHC tensor is antisymmetric in nature, σij=−σji\sigma_{ij}=-\sigma_{ji}. For example, under TT symmetry, Tσxy(k)=−σxy(−k)T\sigma_{xy}(\mathbf{k})=-\sigma_{xy}(-\mathbf{k}). Therefore, the summation in Eq. (8) evaluates to zero, even though the Berry curvature is not zero at every k\mathbf{k}-point. The TT symmetry is satisfied for nonmagnetic materials; a well-known example is MoS2_2. Similarly, PTPT symmetry leads to

PTΩ(k)=−PΩ(−k)=−Ω(k),PT\Omega(\mathbf{k})=-P\Omega(-\mathbf{k})=-\Omega(\mathbf{k}),

which results in zero Berry curvature at every k\mathbf{k}-point [Ω(k)=0][\Omega(\mathbf{k})=0], thereby forbidding the anomalous Hall effect (AHE).

Let us consider the case of MxM_x symmetry. The MxM_x symmetry transforms the Berry curvature as

MxΩ(k)=(Ωx,−Ωy,−Ωz)(−kx,ky,kz).M_x\Omega(\mathbf{k})=(\Omega_x,-\Omega_y,-\Omega_z)(-k_x,k_y,k_z).

Consequently, σzy\sigma_{zy} is the only nonzero component in this case. Alternatively, MxM_x transforms σH\boldsymbol{\sigma}_H as Mx(σzy,σxz,σyx)=(σzy,−σxz,−σyx)M_x(\sigma_{zy},\sigma_{xz},\sigma_{yx})=(\sigma_{zy},-\sigma_{xz},-\sigma_{yx}), which implies that only σyz≠0\sigma_{yz}\ne0. Using a similar procedure for the rest of the symmetry operations, one can build Table I of the main text.

AHE and Néel vector relationship

Figure 4 of the main text shows the functional dependence of the AHE on the Néel vector. Writing the general form of the AHC as σxy=σxy(N)\sigma_{xy}=\sigma_{xy}(N), one can obtain the functional dependence of the AHE on the Néel vector as:

σxy(N)=∑m,n,rλmnrNxmNynNzr,(9)\sigma_{xy}(N)=\sum_{m,n,r}\lambda_{mnr}N_x^mN_y^nN_z^r, \tag*{(9)}

where, λmnr\lambda_{mnr} are the coefficients that determine the contribution of each term. Some terms are forbidden by symmetry. For example, σxy\sigma_{xy} cannot be even in NN as it would violate the Onsager reciprocity relation: σxy(N)=−σxy(−N)\sigma_{xy}(N)=-\sigma_{xy}(-N). The Néel vector is determined by the magnetization of sublattices. Therefore, the Néel vector transforms as a pseudovector under the space group operation, OO. However, it acquires an additional sign (±\pm), depending on whether transformation exchanges opposite (−-) or same (++) spin sublattices (see Figure 1 of main text). Mathematically, this can be expressed as:

ON=±Det(O)D(O)N.(10)ON=\pm Det(O)D(O)N. \tag*{(10)}

where Det⁡(O)\operatorname{Det}(O) is the determinant of the matrix [D(O)][D(O)] for the symmetry operation OO. For instance, if the opposite spin sublattices are connect by Inversion (PP), then PN=(−1)(−1)(−I3×3)NPN = (-1)(-1)(-I_{3\times3})N and therefore PN=−NPN=-N. Also, PP transforms σH\sigma_H as PσH=(−1)(−I3×3)σHP\sigma_H=(-1)(-I_{3\times3})\sigma_H, which mean Pσxy=σxyP\sigma_{xy}=\sigma_{xy}. Therefore, when the opposite spin sublattices are connect by inversion, all of the λmnr\lambda_{mnr} coefficients are symmetry forbidden, suppressing the AHE. This can also be understood from the fact that the system has PTPT ({[C2∥P][C_2\Vert P]} in spin group notation) symmetry and so linear AHE is not allowed (although, higher-order AHE may be allowed). Similar arguments can be used to show that if inversion symmetry connects the same spin sublattices ([E∥P][E\Vert P]), it does not impose any condition on the Hall effect. These observation are consistent with the Table I of the main text.

For the case of 2D altermagnetism (see Sec I of SM), the allowed spin group symmetries are [C2∥C2x,Mx,C4z][C_2\Vert C_2^x,M_x,C_4^z] (which connect opposite spin sublattices) and [E∥E,P,C2z,Mz,C2x,Mx,C3z,C4z][E\Vert E,P,C_2^z,M_z,C_2^x,M_x,C_3^z,C_4^z] (which connect same-spin sublattices). Note that the in-plane direction is chosen along the xx-axis. Other symmetry operations, such as [C2∥S4z][C_2\Vert S_4^z], [E∥S4z][E\Vert S_4^z], and [E∥S3z][E\Vert S_3^z] can be generated from the aforementioned operations, and thus are safely ignored.

Out of 80 layer groups only a few can have spin group symmetry [C2∥C2x,Mx,C4z][C_2\Vert C_2^x,M_x,C_4^z]. The possible layer groups are 2/m2/m, mmmmmm, 4/m4/m, 4/mmm4/mmm, 3m3m, and 6/mmm6/mmm. Note that layer groups containing screw axes and glide planes are ignored, although they can also be included. For example, observations for 2/m2/m should also apply to 21/m2_1/m, 2/b2/b, and 21/b2_1/b. For a particular layer group, symmetry operations will either connect opposite spin sublattices or same-spin sublattices. Operations connecting same- and opposite-spin sublattices are highlighted with superscripts, 1 and 2, respectively, which are added before the symmetry operation. For example, in 24/1m2m2m{}^{2}4/{}^{1}m{}^{2}m{}^{2}m, the fourfold rotation connects opposite spin sublattices, the horizontal mirror connects same-spin sublattices, and the two vertical mirror planes connect opposite spin sublattices. There can be seven nontrivial spin layer groups, namely 22/2m{}^{2}2/{}^{2}m, 2m2m1m{}^{2}m{}^{2}m{}^{1}m, 24/1m{}^{2}4/{}^{1}m, 24/1m2m1m{}^{2}4/{}^{1}m{}^{2}m{}^{1}m, 24/1m2m2m{}^{2}4/{}^{1}m{}^{2}m{}^{2}m, 132m{}^{1}3{}^{2}m, and 16/1m2m2m{}^{1}6/{}^{1}m{}^{2}m{}^{2}m. Therefore, we derive the relationship between σxy\sigma_{xy} and NN for these seven nontrivial spin layer groups in Table II of the main text. Here, we highlight the case of 24/1m2m1m{}^{2}4/{}^{1}m{}^{2}m{}^{1}m (see Table S1). The generators of 24/1m2m1m{}^{2}4/{}^{1}m{}^{2}m{}^{1}m are [E∥E][E\Vert E], [E∥P][E\Vert P], [E∥C2z][E\Vert C_2^z], [E∥C2x][E\Vert C_2^x], and [C2∥C4z][C_2\Vert C_4^z]. The invariant common to all symmetry operations is Nx2Nz−Ny2NzN_x^2N_z-N_y^2N_z. Therefore, for the 24/1m2m1m{}^{2}4/{}^{1}m{}^{2}m{}^{1}m nontrivial spin layer group, we obtain

Symmetry OperationConstraint on σxy(Nx,Ny,Nz)\sigma_{xy}(N_x,N_y,N_z)Symmetry Invariants
[E∥E][E\Vert E]σxy(Nx,Ny,Nz)=σxy(Nx,Ny,Nz)\sigma_{xy}(N_x,N_y,N_z)=\sigma_{xy}(N_x,N_y,N_z)NxmNynNzr (m+n+r=odd)N_x^mN_y^nN_z^r\ (m+n+r=\text{odd})
[E∥P][E\Vert P]σxy(Nx,Ny,Nz)=σxy(Nx,Ny,Nz)\sigma_{xy}(N_x,N_y,N_z)=\sigma_{xy}(N_x,N_y,N_z)NxmNynNzr (m+n+r=odd)N_x^mN_y^nN_z^r\ (m+n+r=\text{odd})
[C2∥C4z][C_2\Vert C_4^z]σxy(Nx,Ny,Nz)=σxy(−Ny,Nx,−Nz)\sigma_{xy}(N_x,N_y,N_z)=\sigma_{xy}(-N_y,N_x,-N_z)Nx2Nz−Ny2NzN_x^2N_z-N_y^2N_z
[E∥C2z][E\Vert C_2^z]σxy(Nx,Ny,Nz)=σxy(−Nx,−Ny,Nz)\sigma_{xy}(N_x,N_y,N_z)=\sigma_{xy}(-N_x,-N_y,N_z)NzN_z, Nx2NzN_x^2N_z, Ny2NzN_y^2N_z
[E∥C2x][E\Vert C_2^x]σxy(Nx,Ny,Nz)=−σxy(−Nx,−Ny,−Nz)=σxy(−Nx,Ny,Nz)\sigma_{xy}(N_x,N_y,N_z)=-\sigma_{xy}(-N_x,-N_y,-N_z)=\sigma_{xy}(-N_x,N_y,N_z)NyN_y, NzN_z, Nx2NyN_x^2N_y, Nx2NzN_x^2N_z, Ny2NzN_y^2N_z, NyNz2N_yN_z^2

Table S1. Summary of invariants under the symmetry operations of 24/1m2m1m{}^{2}4/{}^{1}m{}^{2}m{}^{1}m.

σxy24/1m2m1m=λ201(Nx2Nz−Ny2Nz).\sigma_{xy}^{{}^{2}4/{}^{1}m{}^{2}m{}^{1}m}=\lambda_{201}(N_x^2N_z-N_y^2N_z).

Monolayer MnPSe₃ and MnSe

The opposite spin sublattices in monolayer MnPSe3_3 and MnSe are connected by [C2∥P][C_2\Vert P] symmetry. Therefore, the bands are degenerate throughout the BZ (see Figure S1). Similarly, the bands remain degenerate throughout the BZ with the inclusion of SOC due to PTPT symmetry (see Sec. II). PTPT symmetry also suppresses the AHE in both cases (see Sec. II). Note that we have chosen these two materials as they exhibit tilted Dirac cones near the valence band maximum along the Γ−M\Gamma-M path. These tilted Dirac cones may act as a source of strong Berry curvature when PTPT symmetry is broken (see, for example, Ref. [1]).

Spin-polarized band structures of monolayer MnPSe$_3$ and monolayer MnSe

Figure S1. Spin-polarized band structure of (a) monolayer MnPSe3_3 and (b) monolayer MnSe without SOC. The red and blue curves denote the spin-up and spin-down bands, respectively. The bands are degenerate throughout the BZ due to spin group symmetry [C2∥P][C_2\Vert P]. Similarly, the bands remain degenerate throughout the BZ with the inclusion of SOC due to PTPT symmetry (not shown here; see, for example, Refs. [2, 3]). We make use of PyProcar [4] to plot the band structures.

Altermagnetism in bilayer MnPSe3_3

Figure S2 shows the two possible high symmetry stackings of bilayer MnPSe3_3, named dd-MnPSe3_3 and ii-MnPSe3_3. To understand the altermagnetism in bilayer MnPSe3_3, we plot constant energy surfaces for dd-MnPSe3_3 and ii-MnPSe3_3 in Figure S3. The dd-MnPSe3_3 and ii-MnPSe3_3 exhibit 2 and 6 spin-degenerate nodal lines passing through the Γ\Gamma point, respectively. For dd-MnPSe3_3, this can be understood from the fact that dd-MnPSe3_3 has [C2∥M−3x+y][C_2\Vert M_{-\sqrt{3}x+y}] symmetry, which enforces spin degeneracy along the Γ\Gamma-K1K_{1} path. The [C‾2∥T][C‾2∥M−3x+y]\left[\overline{C}_{2}\mathbin\Vert T\right]\left[\overline{C}_{2}\mathbin\Vert M_{-\sqrt{3}x+y}\right] symmetry enforces spin degeneracy along the other spin-degenerate line. In a similar way, six spin group symmetries for ii-MnPSe3_3—three being [C2∥C2x]\left[C_{2}\mathbin\Vert C_{2}^{x}\right], [C2∥C2x+3y]\left[C_{2}\mathbin\Vert C_{2}^{x+\sqrt{3}y}\right], and [C2∥C2−x+3y]\left[C_{2}\mathbin\Vert C_{2}^{-x+\sqrt{3}y}\right], and three being their combinations with [C‾2∥T]\left[\overline{C}_{2}\mathbin\Vert T\right]—enforce spin degeneracy along the Γ\Gamma-KK and Γ\Gamma-MM directions (six in total). The nonrelativistic dispersions in dd-MnPSe3_3 and ii-MnPSe3_3 are classified as dd-wave and ii-wave altermagnetism, respectively. Similarly, gg-wave altermagnetism exhibits 4 spin-degenerate nodal lines crossing the Γ\Gamma point in the 2D Brillouin zone [5].

Top and side views of bilayer d-MnPSe$_3$ and i-MnPSe$_3$

Figure S2. The top and side views of bilayer dd-MnPSe3_3 [(a) and (b)] and ii-MnPSe3_3 [(c) and (d)]. The blue and yellow spheres denote the Mn atoms from the top and bottom layers, respectively. The green and grey spheres represent P and Se atoms, respectively.

Image not extracted

Figure S3. (a) The Brillouin Zone of the hexagonal unit cell, showing key symmetry points used in the band structure plots. (b) Constant energy contour at −0.23-0.23 eV for dd-MnPSe3_3 and (c) constant energy contour at −0.17-0.17 eV for ii-MnPSe3_3. Red solid lines and blue dashed lines represent spin-up and spin-down bands, respectively. Spin-degenerate nodal lines are highlighted in (a), with blue and red arrows marking the spin-degenerate lines for dd-MnPSe3_3 and ii-MnPSe3_3, respectively.

Origin of AHE in bilayer MnPSe3_3

The Berry curvature shows peaks at the band anticrossings in FMs [6], AFMs [7], and AMs [8]. It is not surprising then that the contribution to AHC in dd-MnPSe3_3 and ii-MnPSe3_3 is highest around their respective band anticrossings (see Figure S4). Therefore, the 2D materials that feature symmetry-protected Dirac and Weyl points and nodal lines may be used to designed AMs with strong anomalous Hall response.

Image not extracted

Figure S4. The spin-polarized band structure (upper panel) and total Berry curvature −Ωz(k)-\Omega_z(\mathbf{k}) (lower panel) for (a) dd-MnPSe3_3 and (b) ii-MnPSe3_3 with SOC and magnetization along the xx-direction. The spin-polarized bands are colored according to the spin expectation along the zz-direction. The Fermi energy is shifted to the energy level at which significant anomalous Hall conductivity is observed, as discussed in the main text (−0.23-0.23 eV and −0.17-0.17 eV below the valence band maximum for dd-MnPSe3_3 and ii-MnPSe3_3, respectively).

AHE in altermagnetic bilayer MnSe

To show that our group theoretic analysis is general and applies to other materials as well, we also performed the calculations on bilayer MnSe. First we obtained the bilayer MnSe by taking the mirror reflection of the lower layer, followed by the in-plane translation of 1/3a+1/3b1/3a + 1/3b. The spin polarized band structure of bilayer MnSe shows the dd-wave altermagnetism [see Figure S5]. The opposite spin sublattices are connected by C2−3x+yC_{2}^{-\sqrt{3}x+y}. The altermagnetic bilayer MnSe show anomalous Hall response similar to dd-MnPSe3_3 [see Figure S5].

Spin polarized bands and anomalous Hall conductivity of altermagnetic bilayer MnSe

Figure S5. (a) Spin polarized bands (without SOC) and (b) anomalous Hall conductivity σxy\sigma_{xy} of altermagnetic bilayer MnSe. The zero Fermi energy corresponds to the valence band maximum.

AHE in weakly ferromagnetic bilayer MnSe

In this section, we discuss how AHE of an altermagnetic bilayer can be distinguished from the weakly ferromagnetic bilayer. For this, we used the bilayer MnSe obtained using the basal plane translation 1/3a+2/3b1/3a + 2/3b and 2/3a+1/3b2/3a + 1/3b. Following Ref [3], we name those stacking arrangements as AB-MnSe and BA-MnSe. The sliding ferroelectricty in bilayer MnSe leads to nonzero net magnetic moment (weak ferromagnetism). A recent study [npj Computational Materials 9 (16), 2023] highlights the weak ferromagnetism in AB-MnSe and BA-MnSe, which is switchable by interlayer sliding from AB to BA stacking. We have plotted the bandstructure of AB-MnSe and BA-MnSe in Figure S6 and Figure S6. The weak ferromagnetism can be understood from the fact that no crystal symmetry connects opposite spin sublattices in these structures. The weak switchable ferromagnetism is evident from the spin polarized band structure in Figure S7 and Figure S7. Next, we obtained the AHC, σxy\sigma_{xy}, of the weakly ferromagnetic bilayer MnSe for different Néel vector orientations. We observe that this weakly ferromagnetic bilayer MnSe shows FM-like Hall response where, the σxy\sigma_{xy} is nonzero when the Néel vector points along zz-direction. Note that σxy\sigma_{xy} is reversed not only upon switching the Néel vector from the zz to −z-z direction, but also by changing the stacking from AB to BA. These FM-like Hall responses in bilayer AFMs have recently gained much attention. For example, FM-type Hall responses have been reported for MnBi2_2Te4_4 through sliding ferroelectricity [9, 10] or external electric field [11].

Spin polarized bands of AB-MnSe and BA-MnSe without SOC

Figure S6. Spin polarized bands of (a) AB-MnSe and (d) BA-MnSe without SOC. The zero Fermi energy corresponds to the valence band maximum.

Anomalous Hall conductivity of AB-MnSe and BA-MnSe as a function of Fermi energy

Figure S7. Anomalous Hall conductivity σxy\sigma_{xy} of (a) AB-MnSe and (d) BA-MnSe as a function of Fermi energy for different orientations of the Néel vector. The zero Fermi energy corresponds to valence band maximum.

References

  1. [1]C. Wang, Y. Gao, and D. Xiao, Intrinsic nonlinear hall effect in antiferromagnetic tetragonal cumnas, Physical Review Letters 127, 277201 (2021).DOI
  2. [2]N. Sivadas, S. Okamoto, and D. Xiao, Gate-controllable magneto-optic kerr effect in layered collinear antiferromagnets, Physical Review Letters 117, 267203 (2016).DOI
  3. [3]K. Liu, X. Ma, S. Xu, Y. Li, and M. Zhao, Tunable sliding ferroelectricity and magnetoelectric coupling in two-dimensional multiferroic mnse materials, npj Computational Materials 9, 16 (2023).
  4. [4]U. Herath, P. Tavazde, X. He, E. Bousquet, S. Singh, F. Muñoz, and A. H. Romero, Pyprocar: A python library for electronic structure pre/post-processing, Computer Physics Communications 251, 107080 (2020).
  5. [5]L. Šmejkal, J. Sinova, and T. Jungwirth, Beyond conventional ferromagnetism and antiferromagnetism: A phase with nonrelativistic spin and crystal rotation symmetry, Physical Review X 12, 031042 (2022).DOI
  6. [6]S. Singh and A. Garcia-Castro, Kagome kmn 3 sb 5 metal: Magnetism, lattice dynamics, and anomalous hall conductivity, Physical Review B 108, 245143 (2023).
  7. [7]H. Chen, Q. Niu, and A. H. MacDonald, Anomalous hall effect arising from noncollinear antiferromagnetism, Physical review letters 112, 017205 (2014).pubmed.ncbi.nlm.nih.gov/24483927
  8. [8]L. Šmejkal, R. González-Hernández, T. Jungwirth, and J. Sinova, Crystal time-reversal symmetry breaking and spontaneous hall effect in collinear antiferromagnets, Science advances 6, eaaz8809 (2020).DOI
  9. [9]R. Peng, T. Zhang, Z. He, Q. Wu, Y. Dai, B. Huang, and Y. Ma, Intrinsic layer-polarized anomalous hall effect in bilayer mn bi 2 te 4, Physical Review B 107, 085411 (2023).DOI
  10. [10]T. Cao, D.-F. Shao, K. Huang, G. Gurung, and E. Y. Tsymbal, Switchable anomalous hall effects in polar-stacked 2d antiferromagnet mnb i2te4, Nano Letters 23, 3781 (2023).DOI
  11. [11]A. Gao, Y.-F. Liu, C. Hu, J.-X. Qiu, C. Tzschaschel, B. Ghosh, S.-C. Ho, D. Bérubé, R. Chen, H. Sun, et al., Layer hall effect in a 2d topological axion antiferromagnet, Nature 595, 521 (2021).

Paper details

vlnc:W_3juz9gth
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