Topological insulator interfaced with ferromagnetic insulators: $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ thin films on magnetite and iron garnets

We report our study about the growth and characterization of $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ thin films on top of ${\mathrm{Y}}_3\mathrm{F}{\mathrm{e}}_5{\mathrm{O}}_{12}$(111), $\mathrm{T}{\mathrm{m}}_3\mathrm{F}{\mathrm{e}}_5{\mathrm{O}}_{12}$(111), $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(111), and $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(100) single-crystal substrates. Using molecular-beam epitaxy, we were able to prepare the topological insulator/ferromagnetic insulator heterostructures with no or minimal chemical reaction at the interface. We observed the anomalous Hall effect on these heterostructures and also a suppression of the weak antilocalization in the magnetoresistance, indicating a topological surface-state gap opening induced by the magnetic proximity effect. However, we did not observe any obvious x-ray magnetic circular dichroism (XMCD) on the Te $M_{45}$ edges. The results suggest that the ferromagnetism induced by the magnetic proximity effect via van der Waals bonding in $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ is too weak to be detected by XMCD, but still can be observed by electrical transport measurements. This is in fact not inconsistent with reported density-functional calculations on the size of the gap opening.

Figure 1

RHEED patterns of $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on magnetic substrates (a) YIG(111); (b) TmIG(111); (c) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(111); (d) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(100), and the corresponding substrates. The thickness of the $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ films is indicated.

Figure 2

XPS measurements of $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on (a) YIG(111); (b) TmIG(111); (c) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(111); (d) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(100). The XPS spectra are shifted to have a better comparison of the peak shapes. “+0.1 eV” means that the measured spectra are shifted by 0.1 eV towards lower binding energy. The black lines are the spectra from the reference sample, namely, 10-QL $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on $\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3$(0001). The red lines are the spectra of 2-QL $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on the respective magnetic oxides. The blue lines are the spectra of thicker $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ (6–10 QLs) on the respective magnetic oxides.

Figure 3

ARPES measurements of $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on (a) YIG(111); (b) TmIG(111); (c) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(111); (d) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(100). The corresponding zoom-in ARPES images in emphasized color are shown on the right side of each spectrum. The dashed black lines indicate the Fermi level, and the red lines indicate the surface states.

Figure 4

Temperature-dependent sheet resistance and magnetoconductance of the $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ films on (a) YIG(111); (b) TmIG(111); (c) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(111); (d) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(100). The insets show the fine structure at low temperature and the Verwey transition temperature for $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$.

Figure 5

(a) Thickness-dependent magnetoconductance of the $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ films on $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(100). The black line is the magnetoconductance of the reference sample, 10-QL $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on $\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3$(0001). The results of the fittings of the HLN formula are denoted for each curve. (b) Magnetoconductance of 6-QL $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on YIG(111). The red line is the fitting of the HLN formula. The orange, blue, and green lines are the contributions from WAL, WL, and $B^2$ of the HLN formula, respectively.

Figure 6

Hall resistance of the $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ films on (a) YIG(111); (b) TmIG(111); (c) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(111); (d) $\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$(100). The left figures of (a)–(c) show the raw data, and the right figures of (a)–(c) emphasize the anomalous Hall effect with the linear background (ordinary Hall effect) subtracted.

Figure 7

XAS and XMCD spectra of the Te $M_{45}$ and Fe $L_{23}$ edges on the Te and $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ films deposited on the magnetic insulators. The measuring temperature and the x-ray incident angle (parallel to magnetic field) with respect to the surface normal are indicated in each figure. $\sigma +$ and $\sigma -$ denote the two XAS spectra taken at opposite applied magnetic fields. The XMCD of Te $M_{45}$ was multiplied by 100 to emphasize the tiny features.

Figure 8

XAS and XMCD spectra of the Te $M_{45}$ and Fe $L_{23}$ edges on (a), (b) the $\mathrm{F}{\mathrm{e}}_x\mathrm{Te}$ film measured at 78 and 300 K, respectively; (c) 1-ML Te deposited on 12-nm Fe; (d) 1-QL $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on 1-ML Te deposited on 12-nm Fe. The measuring temperature and the x-ray incident angle (parallel to magnetic field) with respect to the surface normal are indicated in each figure. $\sigma +$ and $\sigma -$ denote the two XAS spectra taken at opposite applied magnetic field. The XMCD spectra were multiplied to emphasize the features.

Figure 9

(a) XAS of the Te $M_5$ edge on 1-QL $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on YIG(111). The energy was shifted by $\pm 50\mathrm{meV}$ for the black and red curves, respectively. The difference spectrum of the XAS curves with 100-meV energy gap is shown in green. (b) Difference spectra of the shifted Te $M_5$ edges on 1-QL $\mathrm{B}{\mathrm{i}}_2\mathrm{T}{\mathrm{e}}_3$ on YIG(111) for various energy gaps.