Terahertz (THz) spectroscopy is a newly emerging pathway that combines different aspects of basic science and technology, stretching from ultrafast nonlinear optics to microwave photonics. This emerging technology is attracting much attention for the development of advanced imaging1, sensing, medical diagnostics, security, and communication2,3 applications. THz emission spectroscopy4 is based on the fundamental principles of electromagnetism; here, the time-dependent magnetization or polarization of the sample on the subpicosecond timescale acts as a source of electromagnetic radiation with THz frequencies. Due to the large penetration depth and very low photon energy of THz waves, this spectroscopic technique has natural advantages over others, such as being less nondestructive to samples and easily applicable under ambient conditions. As a result of societal demand for high-performance, lightweight, low-cost THz sources, breakthroughs have been made in the development of spintronic-based THz emitters. The emission of THz radiation through the ultrafast spintronic phenomenon occurs in combination with magnetism and photonics. According to ref. 5 THz waves are generated from ultrafast demagnetization6 of ferromagnetic thin films after irradiation with a femtosecond laser pulse, and their intensity is proportional to the second derivative of magnetization with respect to time5 \(\frac{{d}^{2}M}{{{dt}}^{2}}\,\). Subsequently, the spin-to-charge conversion (SCC) mechanism in ferromagnetic/nonmagnetic heterostructures4,7,8,9 was also demonstrated to mediate the THz pulses by the conversion of laser-induced ultrafast spin current to ultrafast charge current9 via the inverse spin-Hall effect (ISHE)10. Due to the efficient SCC at the Rashba interface, this phenomenon is one of the promising sources of spintronic THz emitters. In the search for trending developments, THz emission was discussed due to its large anomalous Hall effect (AHE) via spin-orbit coupling from a single-layer ferromagnetic thin film11. This development has the advantage of disregarding the additional presence of nonmagnetic layers, and the THz emission intensity can be increased by improving the topological nature of the material and the AHE.

The giant AHE and giant anomalous Nernst effect (ANE) are topologically nontrivial phenomena12,13. These phenomena act as operating mechanisms for several cutting-edge applications, such as energy harvesting14, magnetic15, and heat flux sensors16. AHE and ANE have both intrinsic and extrinsic physical origins. For the AHE, the skew scattering and side-jump effect17,18 in the scattering process are known as extrinsic contributions; however, the intrinsic mechanism is explained in terms of an anomalous velocity of the electron in the transverse direction of an applied electric field, which stems from the Berry curvature19 due to the spin-orbit interaction combined with interband mixing20,21. Experimental and theoretical studies suggest that the AHE is enhanced due to the large Berry curvature, as the intrinsic mechanism dominates the extrinsic mechanism. Similarly, an intense Berry curvature22 close to the Fermi energy drastically increases the intrinsic ANE23,24,25.

Weyl semimetals are considered innovators in topological materials and are well known for exhibiting a large AHE and ANE. They have unique electronic structures known as Weyl points and nodes in their momentum spaces, which can be regarded as monopoles or antimonopoles of Berry curvatures. These structures are topologically stable, although they lack any particular symmetry26,27. Weyl fermions, low-energy spin-split massless quasiparticles, are retained near the Weyl nodes and form nondegenerate and linearly dispersed energy band structures. This nontrivial nature of the electronic energy bands helps to enhance the transport properties mentioned above.

According to first-principles calculations, Co2TX (T = transition metal, X = Si, Ge, Sn, Al, Ga) is a prospective magnetic Weyl semimetal due to its high Curie temperature, high spin polarization and large number of Weyl points in the momentum space near the Fermi energy28. In particular, Co2MnGa12,29 (CMG) is a promising magnetic Weyl semimetal that exhibits both a giant AHE30 and ANE12 and a high Curie temperature Tc of ~ 700 K31. In addition to several other advantages, the AHE is now considered an additional origin of spin-based THz emitters11.

Here, we focus on a single-layer CMG Weyl semimetal to study the THz emission, as schematically shown in Fig. 1, and determine the underlying physics. A change in chemical ordering in full Heusler crystal structures can significantly affect the AHE32. Thus, we systematically investigate the transport properties and THz emission of 30-nm-thick CMG thin films as a function of the chemical ordering. The experimental THz setup is provided in the Supplementary Information (S1). Increases in the magnetization, AHE, and THz emission are found with increasing chemical ordering. Subsequently, we experimentally investigate the THz emission with varying thicknesses of the CMG films and explore how the physical origins of the THz emission change with the film thickness with both experiments and theoretical modeling. Finally, the composition of the CMG film is varied to tune the AHE by manipulating the position of the Weyl nodes near the Fermi energy, and we provide direct evidence of the large AHE-influenced THz emission. The typical and popular ferromagnetic alloy CoFeB is selected for comparison of its THz emission intensity with that of the CMG film, and we confirm that the AHE significantly promotes the THz emission. Hence, in this study, we demonstrate a unique THz emitter using a CMG film, and our study can aid in the further exploration of other topological metals for this type of THz emitter.

Fig. 1: Laser-induced THz emission from the Weyl semimetal.
figure 1

Schematic illustration of the THz emitter via a single-layer Weyl semimetal.

Results and discussion

The 30-nm-thick CMG films were deposited by a co-sputtering method using Co and Mn45Ga55 targets in an ultrahigh vacuum (UHV) magnetron sputtering system on a single-crystalline MgO(001) substrate. Inductively coupled plasma (ICP) analyses confirmed that the composition of the deposited CMG films was Co1.01Mn0.45Ga0.54; this is almost an ideal stoichiometric composition of Co:Mn:Ga = 2:1:1. The CMG films were deposited at different deposition temperatures (Ts) varying from room temperature (RT) to 600 °C and the films were annealed at the respective Ts for ~20 min after deposition to improve their crystallinities. The 5-nm-thick MgO and 3-nm-thick Ta layers were sequentially deposited on top of the CMG layer at RT to prevent oxidation as well as degradation of the films over time, particularly during femtosecond laser irradiation in the THz measurements. Note that the Ta layer in the samples serves only as the capping layer, and the effect of the ISHE from the Ta layer on the intensity of the THz emission is negligible (see Supplementary information (S2)).

The crystalline structures of the CMG films were investigated using conventional out-of-plane X-ray diffraction (XRD) with Cu Kα radiation in the θ − 2θ geometry. The magnetic properties were measured with a vibrating sample magnetometer (VSM) with both in-plane and out-of-plane geometries. Transport property measurements were carried out with a physical property measurement system (PPMS) using a standard four-probe method for the samples patterned into the Hall bars. The THz emission was measured with a setup using a standard method33,34. A Ti:sapphire laser and regenerative amplifier with a wavelength of ~800 nm were used to excite and measure the THz wave emitted from the sample with a ZnTe crystal through the electro-optic sampling method33,34 (see also Supplementary information (S1)).

Figure 2a shows the out-of-plane XRD patterns for the CMG films deposited at Ts = RT, 400, 500, and 600 °C. Very clear CMG(002) and CMG(004) XRD diffraction peaks were observed; thus, the films were grown along the [001] orientation on the MgO(001) substrates. The XRD data also confirmed the B2-crystal ordering of the CMG films. To determine the most desirable ordering for the Weyl semimetal, i.e., L21 ordering, we also collected the XRD data on the samples whose film planes were rotated by χ = 54.7° to observe the CMG(111) peak. As shown in Fig. 2b, the CMG(111) peaks were clearly observed for the Ts = 500 and 600 °C samples. The epitaxial growth of all samples was also confirmed from the clear observation of the CMG(202) peaks, as shown in Fig. 2b. The chemical ordering was quantified with the order parameter (S), which could be evaluated from the intensity ratio of the XRD peaks. The order parameters SB2 (SL12) for the B2 (L21) ordering are expressed as the following equations:

$$\it \it {S}_{B2}=\sqrt{\frac{{\left({I}_{002}/{I}_{004}\right)}_{\exp .}}{{\left({I}_{002}/{I}_{004}\right)}_{{\rm{cal}}.}}},$$
$${S}_{L2_1}=\sqrt{\frac{{\left({I}_{111}/{I}_{220}\right)}_{\exp .}}{{\left({I}_{111}/{I}_{220}\right)}_{{\rm{cal}}.}}},$$

where I002, I004, I111, and I220 represent the integrated intensities of the respective XRD peaks, as denoted by the subscripts in both the experimental and calculated data. A linear increase in the B2 order parameter was found with increasing Ts (Fig. 2c). On the other hand, the value of the L21 order parameter was close to zero up to Ts = 400 °C, suddenly began to appear at Ts = 500 °C and reached a maximum at Ts = 600 °C (Fig. 2c). Therefore, an increase in both the B2 and L21 chemical order was achieved with increasing Ts. The value of the saturation magnetization (Ms) monotonically increased with increasing chemical ordering and saturates at approximately Ts = 500 °C, where L21-ordering appeared, as shown in Fig. 2d. The maximum value of Ms was slightly lower than the bulk value of 780 emu/cm312, which was likely due to the existence of some magnetic dead layers at the interfaces of the films. To confirm the large AHE, one of the Weyl properties, the AHE was measured in the CMG films patterned into the Hall bars. The Hall resistivity is empirically given by the following relationship:

$${\rho }_{{yx}}={R}_{0}{\mu }_{0}H+{\rho }_{{\rm AHE}}.$$
Fig. 2: Structural, magnetic, and transport properties.
figure 2

a Out-of-plane XRD patterns of the 30-nm-thick CMG thin films deposited at Ts = RT, 400 °C, 500 °C, and 600 °C. b XRD patterns of the CMG(111) and CMG(202) planes measured with χ ~ 54.7° and 45°, respectively, for all the films. c Order parameters (S) and (d) saturation magnetization (Ms) for the CMG films as a function of Ts. e Magnetic field dependence of the Hall resistivity (ρyx) and (f) the anomalous Hall resistivity (ρAHE) and the anomalous Hall angle (θAHE) as a function of Ts.

Equation (3) is a summation of the ordinary Hall resistivity, which has a linear relationship with respect to the applied magnetic field (μ0H) and the anomalous Hall resistivity (ρAHE), which is proportional to the out-of-plane magnetization. In the Hall geometry, the applied electric current and measured Hall voltage are in the x and y directions, respectively, when the external magnetic field is perpendicular to the film plane, i.e., in the z direction. Figure 2e shows the dependence of the Hall resistivity, ρyx, measured at 300 K with respect to the out-of-plane external magnetic field (μ0H) for the samples prepared at different Ts values. Each curve clearly shows that the respective magnitude of the AHE significantly changes with Ts. The ρAHE is determined by extrapolating the high-field data to the zero-field data. We find that the extracted ρAHE increases with Ts and reaches a maximum at Ts = 500 °C with a value of ρAHE = 15.3 μΩ∙cm, and it slightly decreases to 13.5 μΩ∙cm for the Ts = 600 °C sample, as shown in Fig. 2f. The anomalous Hall angle was evaluated using the following relationship:

$${\theta }_{{\rm AHE}}=\frac{{\rho }_{{yx}}}{{\rho }_{{xx}}},$$

where ρxx denotes the longitudinal resistivity. θAHE increases linearly with Ts and reaches a maximum value of 7.3% for the sample annealed at Ts = 600 °C. Thus, we observe a relationship between the AHE and chemical ordering; here, the AHE is much greater in the L21-ordered samples. These results indicate that the AHE is significantly influenced by the Weyl nature of the L21-ordered films.

Figure 3a shows the Ts dependence of the THz emission intensity at Δt = 0 ps in the CMG films. The measurements were performed with an in-plane bias magnetic field of µ0H = 30 mT, at which the magnetization was saturated for all samples. For all data, the laser was irradiated from the capping layer side of the CMG films. A very clear and monotonic increase in the THz emission intensity is found as a function of Ts (Fig. 3b), and the values are nearly saturated at approximately Ts = 500–600 °C. Thus, the data indicate a relationship among the chemical ordering, AHE, and THz emission intensity. In a single-layer film of a ferromagnetic metal, two different origins are possible for THz emission: ultrafast demagnetization and AHE; these were mentioned in the Introduction section. Thus, two contributions need to be deconvoluted from the THz emission observed in the CMG films, and the influence of both origins (if they exist) need to be estimated via further detailed analysis.

Fig. 3: THz emission from 30-nm-thick CMG films.
figure 3

a THz emission from 30-nm-thick CMG films with different Ts. b THz emission amplitude as a function of Ts. c Schematic illustrations of the geometry of the THz emission; pumping either from the substrate side (top) or from the capping layer side (bottom). d, e Corresponding THz emission from the CMG films with Ts = 600 °C. The open and solid circles represent the data measured in the positive and negative directions of the bias magnetic field, respectively.

Figure 3c illustrates the THz emission from the 30-nm-thick CMG film with different directions of pulsed laser pumping, either to the substrate side or to the capping layer side. In Fig. 3d, e, we show the emitted THz waves measured at a pump fluence of 4.45 mJ/cm2 in the substrate side and capping layer side pumping, respectively. In the figures, we also show the THz emission with different applied bias magnetic fields, the positive (+µ0H) and negative (−µ0H) directions, with solid and open circles, respectively. The polarity of the THz emission is reversed when the applied magnetic field is reversed; this is a prominent indication that the THz emission is of magnetic origin5. However, the polarity of the THz emission does not change when the pumping direction is reversed (Fig. 3d, e). The polarity reversal of the THz emission depends on the pumping direction, and this has been observed both for ferromagnetic/nonmagnetic bilayer films35 and single-layer ferromagnetic films11. For the former case, the sign change of the polarization for the THz emission mainly stems from the reversal of the spin current direction due to the flipping of the sample8; however, in the latter case, the reason behind such reversal can be explained11 in terms of the generation of a nonequilibrium super-diffusive longitudinal charge current inside the magnetic layer. This charge current originates from the backflow of the nonthermal electrons occurring at both interfaces of the ferromagnetic film. The longitudinal charge current (jl) is then converted into the transverse charge current (jt) due to the presence of the AHE, which is expressed as follows:

$${{\boldsymbol{j}}}_{t}={\theta }_{{\rm AHE}}{\boldsymbol{m}}\times {{\boldsymbol{j}}}_{l},$$

where m is the unit vector for the magnetization direction and jl depends on two interfaces; thus, sample flipping, which is a change in the pumping direction, results in a polarity reversal of jl, which is also be observed as the polarity changes in the THz emission. Therefore, the sign reversal of the THz emission due to a change in the pumping direction is an indication that the THz emission is generated via the AHE36. However, in our study, the 30-nm-thick CMG films show no such behavior even though they exhibit a large AHE.

For further investigation, we varied the thickness of the CMG film with the same stacking structure deposited at 600 °C. A deposition temperature of 600 °C was chosen to achieve high chemical ordering for all samples. Figure 4a shows the anomalous Hall resistivity (ρAHE) and anomalous Hall angle (θAHE) as a function of the film thickness (tCMG = 30, 15, 10, 7, and 5 nm). A monotonic decrease in ρAHE and θAHE is observed with decreasing tCMG and is almost zero for tCMG < 5 nm. Hence, the CMG films show a significantly large AHE when the thickness is greater than 5 nm. A similar observation was also reported by ref. 37; in their study, a significant decrease in the anomalous Hall angle of CMG thin films below 20 nm was confirmed by the electronic structure calculations and was caused by a decrease in the majority spin contribution to the Berry curvature. Later, the THz emission from the single-layer CMG film as a function of thickness (tCMG) was measured, as shown in Fig. 4b. All THz emissions were measured with substrate-side pumping. A clear THz emission with a significant amplitude change is observed with different thicknesses. The extracted THz peak intensity is plotted as a function of the film thickness (tCMG) in Fig. 4c. The THz emission intensity first steadily decreases as the thickness is reduced to 10 nm, increases again at 5 nm and further decreases below 5 nm, as shown by the kink at approximately tCMG = 5 nm. When the thickness is reduced below 5 nm, almost negligible THz emission is observed. Next, the origin of the THz emission was investigated as a function of tCMG. Figure 4d–f shows the THz emission measured for the tCMG = 30 nm (d), tCMG = 15 nm (e), and tCMG = 5 nm (f) samples. The open and solid circles denote capping layer side pumping and substrate side pumping, respectively. In Fig. 4d, the 30-nm-thick CMG film did not show any sign reversal of the THz emission, but surprisingly, the 15-nm-thick CMG film (Fig. 4e) showed a clear sign reversal with different pumping directions. A similar sign reversal was found for the films with thicknesses of 10, 7, and 5 nm, as shown in Fig. 4f for the 5-nm-thick films. This drastic transition of the pumping-direction-dependent sign reversal of the THz emission as a function of the thickness led to the further investigation for elucidation on the origin of the THz emission for CMG films with a large AHE, influenced by the Weyl semimetallic nature.

Fig. 4: THz emission from CMG films with different thicknesses.
figure 4

a CMG film thickness dependence of the anomalous Hall resistivity (ρAHE) and anomalous Hall angle (θAHE). b THz emission measured for films with different film thicknesses. c Peak intensity of the THz emission as a function of the CMG film thickness. THz emission measured for (d) 30-nm-, (e) 15-nm-, and (f) 5-nm-thick CMG films with different pumping directions.

Although previous studies have attributed the origin of THz emission from single-layer ferromagnetic films to either the AHE or ultrafast demagnetization, our observations revealed some mixed behavior. Therefore, to study the main origin of the THz emission from the single-layer CMG films, we employed theoretical modeling. Figure 5a shows the THz emission measured for the 15-nm-thick CMG film with capping layer side and substrate-side pumping. Figure 5b shows an illustrative explanation of the theoretical modeling of the THz emission for the CMG films with a thickness that was thinner than the laser light penetration depth; here, both pumping directions are considered: the capping layer side and substrate side. In Fig. 5b, jl1 and jl2 denote the longitudinal charge currents generated at each interface of the CMG film/capping layer and substrate/CMG film, respectively. jAHE and jdemag are the charge currents from two sources, the AHE and ultrafast demagnetization, respectively. The external applied magnetic field is in the z-direction. As shown in Fig. 5b, the asymmetry of jl1 and jl2 causes a sign reversal of the AHE-induced THz emission by flipping the sample; however, the demagnetization-induced THz emission shows no sign reversal. Notably, the AHE-induced THz emission also showed no sign of reversal for thicker films in a previous report11. This was the case for the 30-nm-thick CMG films in our study because the laser light penetration depth is typically ~20 nm.

Fig. 5: THz emission from CMG films due to AHE and ultrafast demagnetization.
figure 5

a THz emission for the 15-nm-thick CMG films when pumped either from the substrate side (top) or from the capping layer side (bottom) with a positive or negative bias magnetic field. b Schematics of the theoretical modeling of the AHE-induced and demagnetization-induced THz emission. c Theoretically predicted and (d) experimentally extracted AHE- and demagnetization-induced THz emission intensities as a function of the CMG film thickness.

Based on the theoretical model, the electric field of the THz emission contributed from the AHE can be written as follows11:

$${E}_{{\rm AHE}}\propto {A}_{{\rm AHE}}\,{\theta }_{{\rm AHE}}\left[{r}_{1}{\lambda }_{1}{e}^{-\frac{{t}_{{\rm CMG}}}{{\lambda }_{\rm T}}} \left(1-{e}^{-\frac{{t}_{{\rm CMG}}}{{\lambda }_{1}}}\right)-{r}_{2}{\lambda }_{2}\left(1-{e}^{-\frac{{t}_{{\rm CMG}}}{{\lambda }_{2}}}\right)\right] {Z}_{{\rm imp}},$$

where Zimp, λ1, and λ2 are defined as follows:

$$\normalsize {Z}_{{\rm imp}}=\frac{{Z}_{0}}{{n}_{{\rm air}}+{n}_{{\rm sub}}+{Z}_{0}\sigma {t}_{{\rm CMG}}}, \\ {\lambda }_{1}={(\lambda }_{\rm e}{\lambda }_{\rm T}/2)/({{\lambda }_{\rm T}-\lambda }_{\rm e}/2), \\ {\lambda }_{2}={(\lambda }_{\rm e}{\lambda }_{\rm T}/2)/({{\lambda }_{\rm T}+\lambda }_{\rm e}/2).$$

Here, r1 and r2 are the reflection coefficients of the two interfaces, λe is the nonequilibrium electron mean free path, λT is the decay length of the THz wave inside the CMG film, Z0 is the impedance of free space, σ is the conductivity for the film, nair and nsub are the refractive indices of the air and substrate, respectively, and AAHE is the cross product of the laser fluence, absorptance of the films, saturation magnetization and average electron velocity11. The top and bottom interfaces of the CMG films are nominally identical, but they differ in terms of atomic and nanostructures. Despite the substrate surface being flat, the surface of the CMG films is rough. This roughness affects the specular reflection of electrons, leading to differences in the reflection coefficients (r1 and r2) at the two interfaces. These differences in the reflection coefficients contribute to inversion asymmetry, as previously discussed11. The THz electric field contributed by ultrafast demagnetization can be expressed as follows38:

$${E}_{{\rm demag}}\propto \frac{\partial M\left(t\right)}{\partial t} {t}_{{\rm CMG}} {Z}_{{\rm imp}}$$

where M(t) is the temporal change in the saturation magnetization of the CMG films due to ultrafast demagnetization. As predicted in Eqs. (6) and (7), the electric field of the THz emission due to the AHE and ultrafast demagnetization exponentially decays and linearly increases with increasing tCMG, respectively. The trends of both contributions are illustrated by the blue line (demagnetization) and red curve (AHE) in Fig. 5c. Therefore, the AHE contribution is expected to dominate in the thinner films. Note that the above discussion may be semiquantitative, and more quantitative discussion requires rigorous theories, considering the wide-range magnetic layer thickness.

Later, the experimental contributions of the EAHE and Edemag were extracted from the experimental data. According to the schematic illustration in Fig. 5b, for the capping layer side pumping, the electric field of the THz emission observed can be expressed as follows:

$${E}_{{\rm{cap}}}\left(+H\right)\propto {E}_{{\rm{demag}}}\left(+H\right)-{E}_{{\rm{AHE}}}\left(+H\right)$$

On the other hand, for substrate-side pumping, the electric field of the THz emission observed is expressed as follows:

$${E}_{{\rm{sub}}}\left(+H\right)\propto {E}_{{\rm{demag}}}\left(+H\right)+{E}_{{\rm{AHE}}}\left(+H\right)$$

From Eqs. (8) and (9), we obtain the electric field of the THz emission due to the AHE and demagnetizations as follows:

$${E}_{{\rm{demag}}}\propto \frac{{E}_{{\rm{sub}}}\left(+H\right)+{E}_{{\rm{ca}}p}\left(+H\right)}{2}$$
$${E}_{{\rm{AHE}}}\propto \frac{{E}_{{\rm{sub}}}\left(+H\right)-{E}_{{\rm{cap}}}\left(+H\right)}{2}$$

Figure 5d shows the EAHE and Edemag extracted from the experiments as a function of tCMG, as indicated by the red and blue solid circles, respectively. The extracted EAHE increases with decreasing tCMG, as predicted by theory, and decreases for tCMG < 3 nm. This decrease is expected because of the decrease in the saturation magnetization and AHE. The extracted Edemag increases with increasing tCMG, as also predicted by theory. By comparing both theoretical and experimental data, we find that the THz emission at tCMG = 30 nm mostly originates from ultrafast demagnetization and that at tCMG ≤ 15 nm originates from the AHE. Note that the extraction of the EAHE and Edemag mentioned above is only precise because the CMG layer thickness is thinner than the laser light penetration depth, as mentioned earlier. Thus, the extracted values of the EAHE (Edemag) for the 30-nm-thick films may be underestimated (overestimated), although this does not alter the deduction of this study.

As mentioned above, the THz emission for the CMG films can be explained by considering the contribution from the AHE and ultrafast demagnetization. These detailed investigations confirm that the AHE has a strong effect on the THz emission in these Weyl semimetals. At the same time, further detailed investigations into the influence of the AHE are needed. For that purpose, we prepared a series of CMG films similarly deposited on single-crystalline MgO(001) substrates with varying compositions. The film thickness was fixed at 10 nm because we focused on the AHE-induced THz emission, and the deposition temperature was selected to be 600 °C. By changing the DC cathode power of the Mn45Ga55 target, the composition of Co1-x (MnGa)1+x was varied. Table 1 provides the valence electron number (Nv) corresponding to the composition (x) of the films. Figure 6a shows the CMG(111) peaks observed in the XRD measurements for the films with different Nvs. The increase in the intensity of the CMG(111) peak with decreasing Nv confirms the high degree of chemical ordering of the films. The additional structural information is shown in the Supplementary Information (S3). The anomalous Hall resistivity (ρAHE) and anomalous Hall angle (θAHE) for the films are shown as functions of Nv in Fig. 6b. Interestingly, ρAHE increases and reaches a maximum value of 12.1 μΩ∙cm at Nv = 27.3 and then decreases with increasing Nv. A similar trend was also reported in a previous study39, in which the highest ANE was also observed for the sample with the largest AHC. As shown in Fig. 6b, the maximum θAHE is also observed for Nv of 27.3–27.7, and the trend of θAHE is similar to that of ρAHE. The values of ρAHE and θAHE for the Nv = 26.8 film are relatively low, which potentially originates from the reduction in saturation magnetization40. Figure 6c displays the THz emission for the films with different Nv. Except for the Nv of 26.8, the films clearly exhibit THz emission. The extracted THz emission intensities for the films are plotted as a function of Nv in Fig. 6d. The THz emission intensity shows a trend similar to that of ρAHE. In particular, the maximum value of the THz emission intensity is observed in the film exhibiting the maximum anomalous Hall coefficient, i.e., the film with Nv = 27.3, which is surprisingly in good accordance. Therefore, the AHE of the CMG films strongly contributes to the THz emission. Note that the THz emission and AHE results for the films with different Nv values were reproducibly obtained (not shown here). Additionally, this AHE-induced THz emission was compared with conventional ISHE-induced THz emission41 (see the details in the Supplementary Information (S4)).

Table 1 Composition ratio (x) and the corresponding valence electron number (Nv) for the fabricated Co1-x (Mn0.5-yGa0.5+y)1+x films with y of 0.0455.
Fig. 6: Structural and transport properites and THz emission for CMG films with different compositions.
figure 6

a XRD patterns of the CMG(111) peaks for the 10-nm-thick films with different valence electron numbers (Nv). b Anomalous Hall resistivity (ρAHE) and anomalous Hall angle (θAHE) as a function of Nv. c THz emission for the films with Nv and (d) their peak amplitudes of the THz emission as a function of Nv.

Finally, we compare the THz emission in the CMG films with that in typical ferromagnetic films, i.e., CoFeB, which is often used in THz emission experiments. The stacking structure of the prepared CoFeB films is similar to that of the CMG films, i.e., single-crystalline MgO(001) substrate/CoFeB (tCoFeB nm)/MgO (5 nm)/Ta (3 nm). For comparison of the efficiency of the THz emission, we show the Hall resistivity as a function of the applied magnetic field in Fig. 7a and the THz emission for the 5-nm-thick CMG and CoFeB films in Fig. 7b. Both the ρyx and THz emission intensities are much greater in the CMG film than in the CoFeB film (see also the details in the Supplementary Information (S5)). Clearly, the large THz emission in CMG films is influenced by the large AHE due to the Weyl semimetal nature.

Fig. 7: Comparison between CoFeB and CMG films.
figure 7

a Magnetic field dependence of the Hall resistivity (ρxy) and (b) THz emission for the 5-nm-thick CoFeB and CMG films.


In summary, for the first time, we experimentally demonstrated that the CMG film is a promising THz emitter influenced by its Weyl semimetallic nature. A strong correlation between the AHE and THz emission was established for highly L21-ordered CMG films. We clearly showed how THz radiation was emitted through the presence of an AHE and ultrafast demagnetization based on the experimental evidence and theoretical modeling considering the super-diffusive backflow current at the interfaces and the transverse current induced by the AHE. We also showed that both the AHE properties and THz emission intensity were influenced by the composition, which resulted from the variation in the Weyl node position near the Fermi energy. To the best of our knowledge, CMG films with an Nv of 27–28 exhibit strong THz emission intensity, which has not been previously reported, even though the highest ANE and AHE39 are known to occur in CMG films with these Nv values. We also demonstrated that the THz emission intensity of the CMG films was much greater than that of the CoFeB films. From an application point of view, our study established that a Weyl semimetal is a promising material for some spintronic devices and can also be used as a THz emitter source42,43. These emitters can be prepared without heavy elements; thus, our study provides a new path to the search for other Weyl materials exhibiting a much larger AHE to attain efficient and prospective THz emitters.

Experimental section

Thin film growth

The CMG films were deposited on single-crystalline MgO(001) substrates using a UHV magnetron sputtering system with a base pressure of ~2 × 10−7 Pa. The Co and Mn45Ga55 targets were used for co-sputtering, and the DC cathode powers for both targets were set at 52 W and 36 W, respectively. The stacking structure of the film was MgO subst./CMG (tCMG nm)/MgO (5 nm)/Ta (3 nm). The Ar gas pressure was 0.33 Pa for the deposition of the CMG film and 0.1 Pa for the deposition of the MgO and Ta films. The substrates were heated at 700 °C to achieve smooth and clean surfaces. Subsequently, the CMG films were deposited at different Ts. The CMG films were postannealed at Ts for 20 min to achieve a smooth surface. Finally, the MgO and Ta layers were deposited as capping layers at RT. The composition of the CMG film was evaluated via ICP analysis. For the fabrication of the films with different compositions, we varied the DC cathode power of the Mn45Ga55 targets for the co-sputtering and deposited a 10-nm-thick CMG film at a Ts of 600 °C via the same stacking process as above.

Measurement of the magnetic and transport properties

The crystal structure of the CMG films was revealed through XRD with a Cu Kα X-ray source. The magnetic properties were measured using VSM. The Hall-bar-shaped samples were used to measure the transport properties. The Hall bars had a width of 42 μm and a length of 84 μm and were patterned from the CMG films through standard photolithography and Ar ion milling. The AHE was measured at 300 K with a current of 100 μA applied along the [110] direction, and a magnetic field was applied perpendicular to the plane for the CMG films using PPMS (Quantum Design Co. Ltd.).

Measurement of the THz emission

THz emission was measured using a custom-built set-up44,45. We used a Ti: sapphire laser and regenerative amplifier with a pulse width of ~120 fs, a wavelength of ~800 nm, and a repetition rate of 1 kHz. A linearly polarized pulsed laser was normally incident on the film surface with a spot size of ~0.9 mm in diameter. The pump pulse intensity was modulated by a mechanical chopper with a modulation frequency of 360 Hz. The THz wave emitted from the film was collimated by a combination of two parabolic mirrors and was focused on a 1-mm-thick ZnTe(110) crystal. Following the probe laser pulse passed through the ZnTe crystal, the THz electric field-induced ellipticity of the probe laser due to the electro-optic effect was analyzed by using a quarter-wave plate, a Wollaston prism, and a balanced photodetector. With the help of the electro-optic sampling method46 and a sweeping delay line, the THz waveform was recorded in the time domain47.