Introduction

The Weyl semimetal (WSM)1,2,3,4,5,6,7 is characterized by the linear band-crossing points, called Weyl points, which exhibits monopole-type structure of the Berry curvature8,9, leading to many exotic properties, such as the Fermi arc surface states1, the chiral anomaly effect10,11, and the anomalous Hall effect (AHE)12,13. This topological phase has been discovered in materials such as TaAs14,15,16,17,18 and MoTe219,20,21,22 recently. In these materials, once the crystal is formed, their positions and energies of Weyl points are usually predetermined by material parameters, including the crystal structure and spin–orbit coupling (SOC), and can hardly be manipulated freely.

A Weyl point is robust in the sense it does not require any symmetry protection except the lattice translation. Inside a mirror plane that prohibits the Berry-curvature monopole, the Weyl point disappears, while 1D nodal lines may emerge due to the mirror symmetry23,24,25. The nodal line displays a π Zak–Berry phase accumulated along a loop circling the nodal line, which induces Shockley-like26,27 surface states (called drum-head surface states28). However, the nodal line exhibits zero Berry curvature in its vicinity and thus does not generate an AHE. If the mirror symmetry is broken, the nodal line gets gapped out and evolves into a pair of Weyl points (e.g., ref. 29). The topology change can be directly probed by the AHE or observed in other topology-induced phenomena. Therefore, the rotation of the magnetization orientation, which sensitively switches the mirror symmetry in a ferromagnet, provides a powerful tool to tune the topological band structure.

Although there have been several reported/predicted magnetic WSMs, few of them are appropriate for tuning the band structure topology via magnetic field. For the antiferromagnetic WSMs Mn3Sn and its sister compound Mn3Ge30,31,32,33,34, in spite of their AHE varying with the rotation of spins within the spin easy plane, it remains unclear how their Weyl points and nodal lines (if they exist) evolve because of the complexity in the band structure. Co3Sn2S2 was reported to be a layered ferromagnetic (FM) WSM35,36,37,38,39, which presents only Weyl points near the Fermi energy. Since its magnetization favors only the out-of-plane direction with the Curie temperature at 177 K, the band tuning by magnetization is challenging, which is similar to the case of a recently reported FM nodal line semimetal Fe3GeTe240. Recently, Co-based Heusler alloy compounds Co2XZ (X=V, Zr, Nb, Ti, Mn, Hf; Z=Si, Ge, Sn, Ga and Al), previously known as half metallic ferromagnets, were predicted to host FM WSM phases41,42,43,44,45,46,47. Since many of these materials are soft ferromagnets with their Curie temperatures far above room temperature, a weak external magnetic field can easily drive their magnetization to rotate in a wide temperature range. Hence, these materials provide ideal platforms to tune the band structure topology.

Although Co2XZ allows for many different element combinations, experimental studies on their possible exotic properties induced by the expected WSM states are sparse, which is possibly due to the difficulty of the single crystal growth of this family of materials. Up-to-date, Co2MnGa is the only member that was reported to have distinct properties associated with the FM WSM state, that is, a giant AHE and anomalous Nernst effect48,49. The band structure calculations show that Co2MnGa as well as other Co2XZ members have both Weyl points and nodal rings, and the giant AHE in these materials is generally believed to originate from Weyl points42,43,46,48,50, with the anomalous Hall conductivity (AHC) being proportional to the separation of Weyl points. In theory, the largest possible Weyl point separation is the Brillouin size, which should induce the quantized AHC, that is, a 3D quantum AHE (QAHE). How can a magnetic WSM reach the 3D QAHE or giant AHE? What is the recipe to find such materials? Answers to these questions are not only of fundamental importance, but also likely leads to technological applications.

In this study, we give clues to these questions through the study of AHE of a Heusler alloy Co2MnAl. This compound has attracted much attention due to its potential applications in spintronics51 and Hall sensors52,53, as well as its recent prediction of FM WSM42. Co2MnAl possesses several structural phases with different disorder types, including L21 (Mn, Al ordered), B2 (Mn, Al disordered), DO3 (Co, Mn disordered), A2 (Co, Mn, Al disordered)52, and only the L21 phase with the cubic space group of \(Fm\overline{3}m\) was predicted to host the FM WSM phase. However, synthesis of the L21 phase has been challenging. All previous studies on this material were based on either polycrystalline or thin film samples with the B2 phase. Our recent success in growing Co2MnAl single crystals with the L21 phase has enabled us to observe its fascinating properties originating from the band topology. We found that this material exhibits very large AHC, up to 1300 Ω−1 cm−1 at room temperature; more noticeably, its room temperature anomalous Hall angle (AHA) ΘH reaches a new record value among all magnetic conductors, with \(\tan\,{\Theta }^{\mathrm{H}}=0.21\). Furthermore, its AHE can be tuned by rotation of the magnetization axis as expected. From the comparison between our experimental results and theoretical calculations, we found that the gapped nodal rings induce large Berry curvature, which is responsible for the observed giant AHE. The consistency between our experiments and theory suggests a new mechanism of creating a large AHE by nodal ring’s gap opening; through this mechanism, an AHC as large as that of a 3D QAHE can be reached. This mechanism is distinct from the previous understanding of the AHE originating from Weyl points for Co2MnGa and similar materials.

Results

Characterization of single crystals

The Co2MnAl single crystals used in this study were grown using floating-zone (FZ) technique (see the “Methods” section). Figure 1b shows an optical image of typical crystals. As shown in Fig. 1a, in the L21 phase, Mn and Al are ordered, which is manifested by the (111) diffraction spot/peak in an electron/X-ray diffraction pattern according to previous studies54. From our electron diffraction analyses, we have indeed observed the (111) diffraction spot in the electron diffraction pattern taken along the [110] zone axis (Fig. 1d), indicating that our Co2MnAl crystals surely have the L21 structure phase. This is further corroborated by scanning transmission electron microscopy (STEM) imaging shown in Fig. 1c, where the periodic, alternating distribution of Mn and Al (left inset to Fig. 1b) can be seen clearly from the atomic intensity line due to Z-contrast. Furthermore, we have also used the high-angle annular dark-field (HAADF) STEM technology to check the sample homogeneity, which confirmed that uniform L21 phase is formed throughout the entire sample.

Fig. 1: Structure and magnetization of Co2MnAl single crystals.
figure 1

a Schematic crystal structure of L21-type Co2MnAl; b the optical graph of Co2MnAl single crystals grown by the optical floating-zone method; c HAADF-STEM image of a selected area taken along the [110] zone axis. The left panel shows alternating intensity due to Z-contrast, revealing the ordering of Mn and Al; the right inset presents a magnified image with the atoms overlaid on top. d The selected area diffraction pattern taken along the [110] zone axis. e The isothermal magnetization data at 2 and 400 K, measured with the magnetic field applied parallel and perpendicular to the (001)/(111) plane, respectively. The insets show the enlargement of the magnetization around zero field.

The magnetic properties of Co2MnAl single crystals were characterized through magnetization M measurements. Figure 1e shows the isothermal magnetization data at 2 and 400 , measured with the magnetic field applied parallel and perpendicular to the (001)/(111) plane, respectively. Field sweep measurements of M reveal small magnetic hysteresis, and the remnant magnetization is ~0.1 and 0.2μB per formula unit, respectively, for B (001) and B (111) at 2 and 400 K (see the insets to Fig. 2e). These data are consistent with the previous report that Co2MnAl is a soft ferromagnet, with the saturation moment Ms of ~4.2 ±  0.2μB per formula54. The small decrease of Ms from 2 to 400 K suggests that its Curie temperature is far above room temperature, which cannot be probed in our superconducting quantum interference device (SQUID) magnetometer. The previously reported Tc for this material is 726 K54. Moreover, our magnetization data also reveal its FM properties are anisotropic and the (111) plane is the easy spin plane.

AHE measurement

We have performed Hall resistivity ρxy measurements on the Co2MnAl samples with the electric current applied to the (001), (010), and (111) planes, respectively. Figure 2a presents the field dependence of ρxy in the 10–300 K temperature ranges for two typical samples with I (001) and I (111). In general, ρxy of an FM conductor can be expressed as ρxy = RHB + 4πRsM, where RHB represents normal Hall contribution due to Lorentz effect (RH, Hall coefficient) and 4πRsM is anomalous Hall contribution \({\rho }_{{xy}}^{A}\). As seen in Fig. 2a, ρxy shows a similar magnetic field dependence as the isothermal magnetization shown in Fig. 1e, indicating that the anomalous Hall contribution (proportional to M) is dominant and the normal contribution is very small and estimated to be 0.3–3% of ρxy in the 2–9 T field range, where M is saturated. \({\rho }_{{xy}}^{A}\) at room temperature shows large values, ~17.9 μΩcm for I (001) and 36.9 μΩcm for I (111). We have also measured field dependence of longitudinal resistivity ρxx at various temperatures for these samples; the representative data from these measurements are presented in Supplementary Fig. 5. From the \({\rho }_{{xy}}^{A}\) obtained after subtracting normal Hall contribution and the ρxx data, we have derived the AHC \({\sigma }_{{yx}}^{A}\) as a function of temperature at 2 T for the I (001) and I (111) samples (Fig. 2b) via tensor conversion, that is, \({\sigma }_{{yx}}^{A}={\rho }_{{xy}}^{A}/({\rho }_{{xy}}^{A\ 2}+{\rho }_{{xx}}^{2})\). Both samples are found to show exceptionally large AHC. At 2 K, \({\sigma }_{{yx}}^{A}\) reaches  ~1600 Ω−1 cm−1 for the I (001) sample. Upon increasing temperature, the AHC decreases slightly and remains up to ~1300 Ω−1 cm−1 at room temperature. For the I (111) sample, the AHC is almost temperature independent and just decreases from ~1250 Ω−1 cm−1 at 2 K to  ~1190 Ω−1 cm−1 at room temperature. Such a large AHC at room temperature in Co2MnAl has never been reported due to the lack of bulk single crystals. The previously reported Hall effect measurements on Co2MnAl were performed on thin film samples with the B2 phase52,53. The giant room temperature AHC revealed in our experiment, as well as its tunability by a weak magnetic field as discussed below, suggests that Co2MnAl is a promising material for device applications such as Hall sensor.

Besides large AHC, Co2MnAl also shows a large AHA, ΘH (\(\tan\,{\Theta }^{\mathrm{H}}={\sigma }_{{yx}}^{A}/{\sigma }_{{xx}}\)), as shown in Fig. 2c. For the I (111) sample, its \(\tan\,{\Theta }^{\mathrm{H}}\) value is as large as 21% even at room temperature, which is a record value among either trivial or topological magnetic conductors, as far as we know. This can be seen in Fig. 2d, where we have compared AHC and AHA between our Co2MnAl samples and those magnetic materials known as having large \({\sigma }_{{yx}}^{A}\) and ΘH. Magnetic WSMs are generally expected to have large AHC and AHA due to the Berry curvature induced by non-trivial band topology. Among reported magnetic WSMs, while some of them indeed exhibit large AHC and AHA, they occur mostly at low temperatures. For instance, the \(\tan\,{\Theta }^{\mathrm{H}}\) of the FM WSM Co3Sn2S2 can also reach 20%, but it can be observed only below 120 K35. The magnetic field induced WSM GdPtBi was also reported to have large AHA, with the largest \(\tan\,{\Theta }^{\mathrm{H}}\) value of ~10% being probed only below 10 K55. As noted above, Heusler compound Co2MnGa has also been reported to have very large AHE48; its AHC reaches ~2000 Ω−1 cm−1 at 2 K, but decreases down to 1000 Ω−1 cm−1 at room temperature. Its room temperature \(\tan\;{\Theta }^{\mathrm{H}}\) is 12%, about half of the largest value we observed in Co2MnAl. Compared to Co2MnGa, Co2MnAl shows a much weaker temperature dependence in AHC and AHA for I (111) (Fig. 2b, c), which is important for device applications.

Fig. 2: The anomalous Hall effect measurement.
figure 2

a Magnetic field-dependent Hall resistivity (ρxy) at various temperatures for two Co2MnAl samples with different orientations. One sample was measured with I (001) plane and B (001) and the other with I (111) and B (111). Left inset: ρxy vs. B at 300 K, with the data for I (001) being offset for clarity. Bottom right inset, Hall coefficient RH as a function of the orientation angle of magnetization for the sample with I (001). The middle right inset shows the experimental setup for the angular-dependent Hall effect measurements. b Temperature dependence of anomalous Hall conductivity (AHC) (\({\sigma }_{{yx}}^{A}\)) for the samples shown in a. c The temperature dependence of \(\tan\,{\Theta }^{\mathrm{H}}={\sigma }_{{yx}}^{A}/{\sigma }_{{xx}}\) at 2 T [ΘH, anomalous Hall angle (AHA)]. d Comparison of AHC and \(\tan\,{\Theta }^{\mathrm{H}}\) between Co2MnAl and other magnetic conductors; “f” refers to thin film materials. Co2MnAl exhibits both large AHA and large AHC in a wide temperature range. In addition to the data collected on the samples shown in a, we have also included the data measured on another two additional Co2MnAl samples with I (001) and I (010) in this figure.The AHC and AHA data of other cited magnetic materials and the related references are given in Supplementary Table 1.

In general, an AHE may either originate from extrinsic mechanism (i.e., skew scattering and side jump), or arises from intrinsic Berry-curvature contribution9, or combined extrinsic and intrinsic contributions8. From the comparison of the temperature dependence of longitudinal resistivity ρxx(T) and AHC \({\sigma }_{{yx}}^{A}(T)\) (Fig. 2c), we find large AHC revealed in our experiments for Co2MnAl is of intrinsic origin. In Supplementary Fig. 6, we show the ρxx(T) data for both I (001) and I (111), from which we observe clear temperature dependence as well as localization behavior below 40 K, indicating the scattering rate varies with temperature. In contrast, \({\sigma }_{{yx}}^{A}\) remains nearly unchanged from several K to 300 K for I (111) despite the scattering rate variation, which is a proof of the intrinsic nature of the AHE according to the theory of AHE in ferromagnets8,9. Furthermore, our theoretical studies presented below also show the intrinsic origin of AHE in Co2MnAl. The \({\sigma }_{{yx}}^{A}\) calculated based on Berry curvature (Supplementary Fig. 1) is close to the experimental value noted above. Additionally, our theory not only explains why \({\sigma }_{{yx}}^{A}(T)\) for I (001) has a slightly negative slope as compared to the I (111) case as shown in Fig. 2c, but also predicts \({\sigma }_{{yx}}^{A}\) can be tuned by the rotation of magnetization, which is also consistent with our experiments, as to be discussed below.

Discussion

To understand the intrinsic AHE of Co2MnAl, we have preformed theoretical calculations on the band structure and AHC of Co2MnAl. The intrinsic AHE originates from the band structure and is robust against disorders and defects. The magnetization direction can modify the AHE, because it determines the symmetry and consequently the band structure of the system. We show the calculated AHC (\({\sigma }_{{yx}}^{A}\)) in Supplementary Fig. 1. It is not surprising to find that the magnetization along [001] and [111] leads to slightly different amplitude in \({\sigma }_{{yx}}^{A}\). Near the charge neutral point, the [111] value is slightly lower than the [001] value and the [111] AHC shows weaker temperature dependence than the [001] case, which is semi-quantitatively consistent with our experiment and is attributed to the band structure anisotropy. \({\sigma }_{{yx}}^{A}\) exhibits different dependence on the chemical potential between the [001] and [111] cases, leading to different temperature dependence of \({\sigma }_{{yx}}^{A}\) as shown in the calculation (Supplementary Fig. 1b) and experiment (Fig. 2b).

We examine the origin of the strong AHE in the band structure by taking the [001] magnetization for example. In an energy window of  −0.2 to  +0.1 eV with respect to the charge neutral point (μ = 0), we find four nodal rings centered at the Z point of the FCC Brillouin zone (see more information in the Supplementary information). We note that these nodal rings are protected by the mirror symmetry of the kz = 0 (kz = 0 and π planes are equivalent in the FCC Brillouin zone), as shown in Fig. 3c and Supplementary Fig. 4b. However, corresponding nodal rings at the kx = 0 and ky = 0 planes are gapped out because of the mirror symmetry breaking caused by the [001] magnetic moment (as shown in Supplementary Fig. 2c). These four nodal rings, which are denoted by #1–4 in the following discussions, are induced by band crossing between the highest two valence bands and lowest two conduction bands (see Fig. 3c). In a previous study (ref. 44), only nodal rings #2 and #3 were investigated in a similar compound Co2MnGa without considering the SOC. Without SOC, the magnetization axis does not pick up a specific direction. Thus, these nodal rings appear in all mirror planes, that is, kx,y,z = 0. As shown in Fig. 3b, four nodal rings in different planes interconnect each other to form topological Hopf nodal links44,56. The SOC leads the magnetic moment to couple to the lattice, thus reducing the symmetry. The preserved nodal rings in the mirror plane do not contribute any Berry curvature (see Supplementary Fig. 2), which is characterized by a π Berry phase accumulated along a loop interlocking the nodal ring. Thus, these four nodal rings do not contribute to the AHE. In contrast, the gapped nodal rings generate large Berry curvature and thus give rise to the huge AHC (see Supplementary Fig. 2) as observed in our experiments. Four peaks of \({\sigma }_{{yx}}^{A}\) in this energy window are dominantly contributed by these four gapped nodal rings (see Supplementary Fig. 3). Near the charge neutral point, AHC exhibits a large value because of the existing peak at μ = 25 meV, which is mainly induced by the gapped nodal ring #3 (see Supplementary Fig. 3), rather than Weyl points. It should be clarified that although Co2MnAl also hosts normal carriers (see Supplementary Fig. 4) besides topological bands, the large AHE is predominantly contributed by the topological carriers.

Fig. 3: The anomalous Hall effect and topological band structure.
figure 3

a The anomalous Hall conductivity \({\sigma }_{{yx}}^{A}\) as rotating the magnetization from the z-axis to y-axis. The experimental and theoretical results are represented by red and black circles, respectively. The solid curve is a cosine profile to guide eyes. b Nodal rings and the first Brillouin zone of Co2MnAl. Without SOC, there are nodal rings in these mirror planes, kx,y,z = 0 or π. For example, there are four nodal rings that are labeled as #1–4 in c, centered at the Z point of the FCC Brillouin zone. When magnetization align along the [001] direction by SOC, all nodal rings are gapped out because of the mirror symmetry breaking, except those in the kz = 0 or π plane. The preserved nodal rings and related band structure are shown in c. The band dispersion, in which the nodal ring crossing points are indicated by blue dots, is along the black dotted line in the upper panel. df The Berry-curvature (Ωyx) distribution in the kx = 0 plane centered at Γ for different magnetization angle (θ). Nodal rings (black) are gapped out in this plane. Near the charge neutral point, the Berry curvature is mainly contributed by the gapped nodal ring #3. When θ = 45°, nodal rings #1 and #4 evolve into Weyl points (filled black circles in e), while nodal rings #2 and #3 are fully gapped. g The nodal ring #3 at four corners can be treated as the reconstruction of a large ring, which is illustrated by a green dashed circle, centered at Γ.

We can control the AHE response by rotating the magnetization axis, when fixing the current and Hall voltage probes on the (001) plane (see the right inset of Fig. 2a). In the experiment, we observed a \(\cos (\theta )\)-like oscillation of \({\sigma }_{{yx}}^{A}\) with respect to the magnetization angle, θ. As shown in Fig. 3a, the experiment is consistent with our theoretical calculations. As the magnetization changes from [001] to [010], nodal rings appear in the the ky = 0 plane, while they disappear in kx,z = 0 planes. In between [001] and [010], #3 nodal rings are fully gapped in all three planes, although some other nodal rings (e.g., #1 and #4) evolve into Weyl points (Fig. 3e). As a consequence, the induced Berry curvature Ωyx evolves dramatically from [001] to [010], as shown in Fig. 3d, f, leading to the oscillating \({\sigma }_{{yx}}^{A}\) as seen in experiments. The nearly cosine-like profile is consistent with the fact that Co2MnAl exhibits weak anisotropy in the magnetization orientation in experiment. We should point out that the present \(\cos (\theta )\)-type AHE is different from the known planar Hall effect, which has a sin(2θ) dependence57. The planar Hall effect appears in an FM metal where the magnetization remains unchanged as rotating an in-plane magnetic field, or a nonmagnetic WSM where the chiral anomaly plays a role58,59. The gap opening of the various nodal lines and the associated band structure reconstruction driven by the rotation of magnetization is manifested by the angular dependence of Hall coefficient RH. In Supplementary Fig. 7, we present the ρxy data collected under various magnetization orientations between [001] and [010]; RH can be extracted from the linear fit in the field range where magnetization is saturated for each orientation. We found that RH varies remarkably as the magnetization is rotated from [001] to [011], with a sign change near 20°, as shown in the right inset of Fig. 2a. Moreover, RH for M [111] is almost four times of that for M [001], see the left inset of Fig. 2a. These observations, together with the angular dependence of \({\sigma }_{{yx}}^{A}\) shown in Fig. 3a, is consistent with the theoretical picture of the band structure and topology tuning by the rotation of magnetization in Co2MnAl discussed above.

In principle, a simple nodal ring can evolve into a pair of Weyl points when the magnetization breaks the corresponding mirror symmetry, as illustrated in Fig. 4a. The induced AHC is proportional to the Weyl point separation12,13. This scenario is seemingly consistent with our observation on nodal rings #1 and #4. However, we do not observe the existence of Weyl points for nodal ring #3, which is the main source of the AHE observed in the experiment, when rotating the magnetization axis.

Fig. 4: Schematic transition between nodal rings and Weyl points.
figure 4

a The nodal ring (solid ring) is protected by the in-plane mirror symmetry where the spin (arrow) points out of plane, in the upper panel. If the spin rotates to break the mirror symmetry, the nodal ring gets gapped out (dashed ring), giving rise to a pair of Weyl points (blue spheres). The induced AHE is proportional to the separation of the Weyl points, that is, the diameter of the nodal ring (kd) in a form \(\frac{{e}^{2}}{h}\frac{{k}_{\mathrm{d}}}{2\pi a}\), where a is the lattice parameter. b The nodal ring is as large as the Brillouin zone size, as a critical point. The resultant Weyl point is pushed to the zone boundary and meets another Weyl point with opposite chirality from the second zone. The induced AHE is quantized to \(\frac{{e}^{2}}{h}\frac{1}{a}\). c The nodal ring is larger than the Brillouin zone size, as an open nodal ring. Then, the gapped ring does not induce Weyl points in the open direction, generating a 3D quantized Hall conductivity, \(\frac{{e}^{2}}{h}\frac{1}{a}\).

Actually the missing of Weyl points for nodal ring #3 can explain the large magnitude of the AHC observed in Co2MnAl. Take the [001] magnetization as an example. The theoretical value of \({\sigma }_{{yx}}^{A} \sim 1400\,{\Omega }^{-1}\,{\mathrm{cm}}^{-1}\) is as large as the AHC of a 3D QAHE60, \(\frac{2{e}^{2}}{h}\frac{1}{a}=1347\,{\Omega }^{-1}\,{\mathrm{cm}}^{-1}\), where a takes the experimental value 5.75 Å. We note that this is not coincidence. Although this nodal ring is centered at the Z point, it emerges from a large nodal ring that is centered at Γ. Because it even crosses the Brillouin zone boundary, the Γ-centered ring gets reconstructed to smaller rings at the zone boundary (see Fig. 3g). Although smaller nodal rings generate Weyl points, a reconstructed nodal ring does not generate Weyl points (illustrated in Fig. 4c). This is because a Weyl point, if it exists, must annihilate with another one from the neighboring Brillouin zone. As a consequence, the band structure has a gap and the induced AHC has a quantized value \(\frac{{e}^{2}}{h}\frac{1}{a}\). In our material, there are two groups of gapped nodal rings on the kx,y = 0 planes. Therefore, the ideal AHC is \(\frac{2{e}^{2}}{h}\frac{1}{a}\). Because the nodal ring cannot be fully gapped out in the Fermi energy and there are also contribution for other bands, we do not expect a true 3D QAHE here. We note that the same scenario can also be applied to Co2MnGa. In addition, if the Γ-centered ring was smaller than the Brillouin zone size in Co2MnAl, it could generate Weyl points but smaller AHC than the quantized value. This is actually the case of Heusler compounds Fe2MnX (X =  P, As, Sb)46, Co2VGa45, and Co2TiX (X = Si, Ge, Sn)43. We can rationalize these full Heusler magnets in the same framework here.

Conclusions

Based on the understanding of Co2MnAl, we can obtain useful insights to design magnetic materials with strong AHE. To generate large AHC, we need multiple large nodal rings that are larger than the Brillouin zone size. To host multiple nodal rings (in the absence of SOC), we prefer to have multiple mirror planes in the crystal structure, which usually indicates high-index space groups. In general, highly symmetric magnetic materials with strong SOC are an optimal choice, not only for the AHE, but also for other Berry-curvature-induced phenomena like the anomalous thermal Hall effect and anomalous Nernst effect. Similarly, nonmagnetic materials with many mirror planes and strong SOC are ideal for the spin Hall effect, as pointed out in ref. 61.

Our theory and experiment together also demonstrate that the form of the gapless nodes (Weyl points or nodal rings), as well as their locations and velocities, all strongly rely on the magnetization axis. Therefore, the magnetization can potentially be utilized to engineer an artificial gauge field62,63 and spacetime geometric structure64,65 of Weyl fermions. On the other hand, Weyl fermions can also mediate a strong coupling between magnetization dynamics and electromagnetic fields63, thus allowing for the electric control of magnetic dynamics in this system. Thus, Co2MnAl provides an ideal intrinsically magnetic platform to explore the interplay between magnetic dynamics and topological physics for the development of a new generation of spintronic devices.

Methods

Experimental

Co2MnAl single crystals were grown using FZ technique. The material rods used for the FZ growth were made using an induction furnace. The crystals were annealed at 1550 K for 1 week, followed by annealing at 873 K for another week. The powder X-ray diffraction measurements confirmed that the Co2MnAl single crystals grown using the above method possesses a cubic structure. Since Co2MnAl has multiple structural phases with different types of disorders as mentioned in the main text and only the L21 phase is predicted to host an FM WSM phase, we performed detailed TEM analyses to identify its structure phase. The TEM sample was prepared using a Thermo Fisher Helios NanoLab Dual-Beam focused ion beam system. The HAADF-STEM images are taken with the Thermo Fisher Titan S/TEM equipped with a spherical aberration corrector, and it was operating at 300 kV accelerating voltage with a probe convergence angle of 30 mrad. Magnetization properties were measured using a SQUID (VSM, Quantum Design). The electrical transport properties were measured using a standard four probe method in a physical property measurement system (Quantum Design).

Ab initio calculation

We have performed density-functional theory (DFT) calculations with a full-potential local-orbital minimum-basis (FPLO)66 code to calculate electronic structure. The exchange and correlation energies are considered in the generalized gradient approximation by Perdew–Burke–Enzerhof scheme67. SOC is included. From the DFT electronic structure, we have projected the Bloch wave function into Wannier functions (Co-3d, Mn-3d, and Al-3p orbitals), to construct an effective Hamiltonian (\(\hat{H}\)) for the bulk material.

We have evaluated the AHC (\({\sigma }_{{yx}}^{A}\)) and Berry curvature (Ωyx) by the Kubo-formula approach in the linear response scheme9,

$${\sigma }_{ij}^{A}(\mu )=-\frac{{{\rm{e}}}^{2}}{\hslash }\int\ {\mathrm{d}}\xi \frac{\partial \ f(\xi -\mu )}{-\partial \xi } \int_{BZ}\frac{{\mathrm{d}}{\bf{k}}}{{(2\pi )}^{3}}\mathop {\sum}\limits_{{\epsilon }_{n}\,{<}\,\mu }{\Omega }_{ij}^{n}({\bf{k}}),$$
(1)
$${\Omega }_{ij}^{n}({\bf{k}})={{i}}\mathop {\sum}\limits_{m\ne n}\frac{\langle n| {\hat{v}}_{i}| m\rangle \langle n| {\hat{v}}_{j}| m\rangle -(i\leftrightarrow j)}{{\left({\epsilon }_{n}({\bf{k}})-{\epsilon }_{m}{\bf{(k)}}\right)}^{2}}.$$
(2)

Here ϵn is the eigenvalue of the \(\left|n\right\rangle\) eigenstate, and \({\hat{v}}_{i}=\frac{{\mathrm{d}}\hat{H}}{\hslash {\mathrm{d}}{k}_{i}}\) (i = xyz) is the velocity operator, μ is Fermi level of the system, and f(ξ − μ) is the Fermi–Dirac distribution. A k-point of grid of 300 × 300 × 300 is used for the numerical integration in Eq. (1).