Optical Readout of the Néel Vector in the Metallic Antiferromagnet ${\mathrm{Mn}}_2\mathrm{Au}$

Metallic antiferromagnets with broken inversion symmetry on the two sublattices, strong spin-orbit coupling, and high Néel temperatures offer alternative opportunities for applications in spintronics. Especially ${\mathrm{Mn}}_2\mathrm{Au}$, with a high Néel temperature and high conductivity, is particularly interesting for real-world applications. Here, manipulation of the orientation of the staggered magnetization, (i.e., the Néel vector) by current pulses was recently demonstrated, with the readout limited to studies of anisotropic magnetoresistance or x-ray magnetic linear dichroism. Here we report on the in-plane reflectivity anisotropy of ${\mathrm{Mn}}_2\mathrm{Au}$(001) films, which are Néel vector aligned in pulsed magnetic fields. In the near-infrared region, the anisotropy is approximately 0.6%, with higher reflectivity for the light polarized along the Néel vector. The observed magnetic linear dichroism is about 4 times larger than the anisotropic magnetoresistance. This suggests the dichroism in ${\mathrm{Mn}}_2\mathrm{Au}$ is a result of the strong spin-orbit interactions giving rise to anisotropy of interband optical transitions, which is in line with recent studies of electronic band structure. The considerable magnetic linear dichroism in the near-infrared region could be used for ultrafast optical readout of the Néel vector in ${\mathrm{Mn}}_2\mathrm{Au}$.

Figure 1

The sample layout and the XMLD PEEM images of samples A–C, shown in (a)–(c), respectively. ${\mathrm{Mn}}_2\mathrm{Au}$ grows epitaxially with the [110] and [$1\overline{1}0$] axes parallel to the substrate edges, which are along the $[010{]}_s$ ($x$) and $[211{]}_s$ ($y$) axes of the r-cut sapphire. Sample A is as grown, while samples B and C are orthogonally polarized in the 60-T pulsed magnetic field. The directions of the applied magnetic field and the corresponding Néel-vector directions are shown in the center, together with the scale of the XMLD PEEM images. The contrast shows the two types of magnetic domains with the Néel vector along the two orthogonal easy axes: [110] and [$1\overline{1}0$]. In the as-prepared sample (A), roughly equal volume fractions of both types of domains are observed, while the aligned samples (B and C) display mostly one type of domain. (d) The unit cell of ${\mathrm{Mn}}_2\mathrm{Au}$, with spin orientations in adjacent layers shown by arrows.

Figure 2

(a) The experimental setup (top view). The laser diode (LD) is vertically polarized ($s$-polarized). The polarization is modulated by the PEM operating at 50 kHz. The distance between the LD and the sample is about 2.5 m, while the photodiode (PD) is mounted about 3 cm from the laser output, giving an incidence angle of $\alpha /2\approx {0.34}^{\circ }$. (b) Time evolution of the light intensity on the PD when a polarizer is placed in front of the PD selecting the horizontal ($p$-polarized) signal (the modulation period is $10\mu \mathrm{s}$). The blue arrows represent the polarization states at given times, varying between vertical ($s$) and horizontal ($p$) polarizations, which are along the [110] and [$1\overline{1}0$] crystalline axes of ${\mathrm{Mn}}_2\mathrm{Au}$. C, circularly polarized light; P, $p$-polarized light; S, $s$-polarized light.

Figure 3

Results of reflectivity anisotropy studies of ${\mathrm{Mn}}_2\mathrm{Au}$ films with different orientations of the magnetic order. (a) The raw reflectivity modulation signals from all three samples, oriented in the same growth direction (raw data obtained on samples rotated by ${90}^{\circ }$ are not shown). The time window here is $20\mu \mathrm{s}$. (b)–(d) The differential signals disentangling the MLD and structural contributions to reflectivity variation according to Eqs. (3) and (4). Here the traces are folded into the 10-$\mu \mathrm{s}$ time window to emphasize the lack of circular dichroism, which should have a 50-kHz component. The modulation in (b),(c) is a result of the MLD, while the modulation in (d) reflects the structural component of dichroism. The modulation of reflectivity is recorded with a digital oscilloscope in the ac coupling mode; thus, the zero baseline is automatically determined by the oscilloscope as the mean value of the periodic signal. Because of the asymmetry of the signal, the zero level does not correspond to the arithmetic average of the maximum and minimum. The asymmetry is a result of PEM operation, where the flat regions [here the light is $p$-polarized; see Fig. 2] correspond to the maximal retardation of the photoelastic modulator. LC, left-circularly-polarized light; P, p-polarized light; S, s-polarized light; RC, right-circularly-polarized light.

Figure 4

The sample layout in dichroism measurements for the two different orientations of the sample (left and right) with respect to the laser-beam directions (red lines). Here $x$ and $y$ correspond to the [110] and [$1\overline{1}0$] directions of ${\mathrm{Mn}}_2\mathrm{Au}$, while $z$ is parallel to the $c$ axis, [001]. The tilt of the $c$ axis is highly exaggerated.