Nanoscale electronic variations in altermagnetic $α$-MnTe

Abstract

Altermagnets exhibit spin-split electronic bands like ferromagnets, yet they are magnetically compensated like collinear antiferromagnets. Altermagnets thus combine the benefits of ferromagnets and antiferromagnets, opening routes to tailored materials and applications. However, reported bulk signatures such as the anomalous Hall response in candidate altermagnets have been inconsistent across samples, suggesting that inhomogeneity may affect their functionality in electronic devices. Here, we use low-temperature scanning tunneling microscopy and spectroscopy to map the local electronic landscape of $α$-MnTe on atomically flat cleaved single crystals. We resolve two distinct electronic regions. In Region A, the chemical potential lies near the valence-band edge and varies by $\sim$100 meV on the nanometer length scale. In Region B, the chemical potential lies near the middle of a wider band gap. We further identify an incommensurate charge modulation with periodicity (2.5$\pm$0.1)$a$, observed exclusively in Region A. Our work establishes that $α$-MnTe can exhibit significant electronic non-uniformity, suggesting that nanoscale characterization is essential for its reliable use in electronic applications.

Figures (4)

• Figure 1: $\alpha$-MnTe crystal. (a) Crystal structure of $\alpha$-MnTe (space group $P6_3/mmc$) with lattice constants labeled moseley2022PRM. Mn moments, represented by red and blue arrows, are ferromagnetically aligned within each $ab$ plane and antiferromagnetically aligned between adjacent layers. (b) Photograph of the cleaved surface of an $\alpha$-MnTe crystal, showing fractured regions of flat surfaces. The inset shows the crystal residue on the cleaving post. (c) Constant-current STM topograph of cleaved $\alpha$-MnTe showing atomic-scale contrast. In the NiAs structure, Mn and Te atoms share the same in-plane lattice periodicity, thus the atomic species cannot be unambiguously assigned from topography alone. Sample bias $V_{\mathrm{s}}=-0.8$ V, setpoint current $I_{\mathrm{s}}=250$ pA. (d) FT of the topograph in (c) with green circles marking the Bragg peaks.
• Figure 2: Long-range electronic variation. (a,b) Constant-current STM topographs acquired in the same field of view at negative and positive sample bias, respectively. Setpoints: (a) $V_{\mathrm{s}}=-0.8$ V, $I_{\mathrm{s}}=400$ pA; (b) $V_{\mathrm{s}}=+1.5$ V, $I_{\mathrm{s}}=200$ pA. (c) $dI/dV$ map at $-0.5$ V from a $64\times64$ spectroscopy grid acquired over the same region as (a,b). (d) Spatially averaged $dI/dV$ spectra for the 10 bins defined by their value at sample bias $-0.5$ V. Spectra are smoothed along the energy axis using a 32 mV moving-average (boxcar) filter. (e) Zoom-in of (d) over the range in the dashed brown box, showing a small in-gap state between 0 and 0.6 V. Spectra were slightly shifted upward to account for a small negative $dI/dV$ contribution. (f) Angle-averaged cross-correlation coefficient $\alpha(r)$ between the $dI/dV$ map at $-0.5$ V [panel (c)] and the low-pass-filtered topographs at $V_{\mathrm{s}}=-0.8$ V (blue) and $V_{\mathrm{s}}=+1.5$ V (orange), as well as between the two low-pass-filtered topographs (green). Spectroscopy grid setpoint: $V_{\mathrm{s}}=-1.3$ V, $I_{\mathrm{s}}=200$ pA, with lock-in modulation (zero-to-peak) $V_{\mathrm{exc}}=6$ mV.
• Figure 3: Incommensurate charge modulation. (a,b) FT of the topographs shown in Fig. \ref{['fig2']}(a,b), respectively. At negative bias $V_{\mathrm{s}}=-0.8$ V, we observe only the Bragg peaks (green circles). At positive bias $V_{\mathrm{s}}=+1.5$ V, additional broad peaks ($Q_\text{CM}$, denoted by the dashed blue circle) appear along the Bragg-peak directions $q_1$, $q_2$ and $q_3$, consistent with a disordered charge modulation. We apply a Hanning window before the FT to reduce edge artifacts. (c) Left: linecut along $q_2$ of (a) with $V_{\mathrm{s}}=-0.8$ V. Right: linecut along $q_2$ of (b) with $V_{\mathrm{s}}=+1.5$ V. Both are averaged over a 10-pixel-wide strip. Shaded areas show the fits to a sum of Lorentzians for the central $q=0$ peak (gray) and the $Q_\text{CM}$ peak (blue) that appears only for the positive-bias linecut. Green boxes show a zoomed-in view around the Bragg peak positions, which were determined by additional Lorentzian components. The horizontal axis is normalized by $|\mathbf{G}|=Q_\mathrm{Bragg}$. (d) High-pass-filtered topograph of Fig. \ref{['fig2']}(b) with long-range variation removed to enhance the visibility of the charge modulation at $V_{\mathrm{s}} = 1.5$ V. (e) 2D autocorrelation of the filtered topograph in (d). (f) Linecuts of the autocorrelation function $\alpha$ along the three directions marked by the dashed lines in (e), averaged over a 10-pixel-wide strip. The profiles illustrate the rapid decay of $\alpha$ from the center, indicating a short correlation length of the charge modulation.
• Figure 4: Two types of regions with distinct electronic features. (a) Schematic of MnTe sample surface with different electronic features. Purple (Region A) denotes areas that host disordered charge modulation, orange (Region B) denotes areas without charge modulation, and the green square encompasses a field of view of mixed A and B regions. (b,c) Representative constant-current topographs acquired in A and B regions, respectively. (d,e) Spatially averaged $dI/dV$ spectra acquired in the fields of view shown in (b,c), respectively. (f) Topograph of a mixed A+B field of view at $V_{\mathrm{s}}=+1.5$ V; dashed contours trace the region boundaries as a guide to the eye. (g) Topograph of the same field of view as (f) at $V_{\mathrm{s}}=-1.3$ V showing continuous atomic resolution and no charge modulation apparent in either A or B regions at negative bias. (h) $dI/dV$ map at $-0.8$ V in the same field of view as (f). (i) FT of A (right) and B (left) regions of the field of view (f), masked based on their $dI/dV$ at $-0.8$ V. (j) Histogram of $dI/dV$ values at $V_\mathrm{ref}=-0.8$ V from the spectroscopy grid in (h), showing a bimodal distribution associated with A-type and B-type regions. (k) Averaged $dI/dV$ spectra for the two populations identified from the bimodal distribution in (j). Setpoints: (b) $V_{\mathrm{s}}=+1.5$ V, $I_{\mathrm{s}} = 200$ pA; (c) $V_{\mathrm{s}}=+1.2$ V, $I_{\mathrm{s}} = 300$ pA; (d) $V_{\mathrm{s}}=-1.3$ V, $I_{\mathrm{s}}=200$ pA, $V_{\mathrm{exc}}=6$ mV (zero-to-peak); (e) $V_{\mathrm{s}}=+1.2$ V, $I_{\mathrm{s}}=300$ pA, $V_{\mathrm{exc}}=6$ mV (zero-to-peak); (f,i) $V_{\mathrm{s}}=+1.5$ V, $I_{\mathrm{s}} = 100$ pA; (g) $V_{\mathrm{s}}=-1.3$ V, $I_{\mathrm{s}} = 100$ pA; (h,j,k) $V_{\mathrm{s}}=-1.3$ V, $I_{\mathrm{s}}=100$ pA; $V_{\mathrm{exc}}=6$ mV (zero-to-peak). Data for Region A and the mixed Region A+B are acquired on one crystal; data for Region B are acquired on a different crystal.