betaFeedback

WorksPhysics and Astronomy

Digitally-alloyed magnetic topological insulator nanostructures grown via stencil mask molecular beam epitaxy

Abstract

Magnetic topological insulator nanostructures are a promising platform for the realization of fault-tolerant quantum computation. For this, such nanostructures need to be embedded into the necessary superconducting architectures. In this study, we demonstrate a method for the growth of magnetic topological insulator nanostructures via molecular beam epitaxy, utilising digital alloying and defining their shapes in-situ by stencil mask technology. Cr-doped (Biy_ySb1−y_{1-y})2_2Te3_3 micro- and nanometer-sized Hall bars are fabricated and the magneto-transport properties for various Cr doping amounts and different Hall bar dimensions are investigated at low temperatures. We find indications that the Fermi energy is located inside the bulk band gap, while disorder and charge puddles affect the transport behaviour. Furthermore, we investigate the magnetic properties and discuss the transition from ferromagnetic to paramagnetic behaviour.

Introduction

Magnetic topological insulators (MTIs) are a promising material platform for realizing fault-tolerant quantum computation utilizing Majorana zero modes [1, 2, 3]. To achieve this, the chiral edge channels of an MTI nanoribbon need to be proximitized with an s-wave superconductor [4, 5]. There are already some reported cases where superconducting correlations are present in antiferromagnetic [6, 7] and ferromagnetic MTIs [8, 9], but theoretical studies motivate to explore larger magnetic doping for superconducting applications [4, 5, 10, 11]. A prominent material is Cr-doped (BiySb1−y)2Te3(\mathrm{Bi}_y\mathrm{Sb}_{1-y})_2\mathrm{Te}_3. Previously, it has been studied how the ferromagnetic and magneto-transport properties of this type of MTI change as a result of varying the Cr content [12, 13, 14]. Therefore, in a previous work we investigated low Cr-doped magnetic topological insulator thin films grown via molecular beam epitaxy (MBE) and showed that even for weak antiferromagnetization, a pronounced anomalous Hall effect is observed [15]. This means that we are able to tune the MTI into a regime that is very promising for superconducting applications [4, 5]. A challenge of embedding MTI nanoribbons in superconducting architectures is the need for very clean interfaces. For regular topological insulator nanostructures grown via MBE, a complete in-situ process has proven to be superior to ex-situ processes [16]. It also has been reported that electron beam lithography as a tool to create nanostructures can cause damage to the MTI material [17, 18] [19].

In order to tackle both problems at once, we developed a fully in-situ process scheme to fabricate MTI nanostructures by using MBE and stencil-masks. In order to demonstrate the feasibility of this processing scheme, Hall bars are fabricated in varying dimensions ranging from micro- to nanometer-sized devices. Furthermore, we investigate how Cr content affects the magnetic and electric properties at low temperatures, and discuss how narrowing down the width of the Hall bars changes these properties. The latter is particularly interesting because narrow MTI nanoribbons play a central role in proposed Majorana architectures.

Methods

Digital alloying

Compared to the conventional MBE growth, where all materials are co-evaporated, for our employed digital alloying approach, all relevant materials are deposited sequentially. This technique has been previously used to precisely tune the composition of compound semiconductors [20, 21, 22] or to grow conventional topological insulator films from Bi2_2Te3_3 and Sb2_2Te3_3 [23, 24] and thin films of the intrinsic MTI MnBi2_2Te4_4 [25]. In order to utilize this approach for our MTI micro- and nanostructures, stencil masks are fabricated onto the Si(111) substrate using SiO2_2 and Si3_3N4_4 layers similar to the work by Schüffelgen et al. [16]. These stencil masks define the shapes of the MTI nanostructures, in our case Hall bar devices, during growth. The sequence of the growth process in the MBE chamber is illustrated in the diagram given in Figure 1(a). After the substrate has been loaded into the MBE chamber it is degassed at TdegasT_{\mathrm{degas}} (650°C) to remove the hydrogen bonds formed during the previous hydrofluoric acid cleaning step from the Si(111) surface. The subsequent layer growth happens in cycles. Each cycle contains the following steps: First, the substrate is heated to TgrowthT_{\mathrm{growth}} (100 - 110°C). The cell of the first element, in this case Bi, is oriented at a specific azimuthal angle φ\varphi relative to the substrate. Then, Bi is evaporated from this direction. Afterwards, the substrate is rotated so that the Cr cell has the same effective orientation to the substrate as the Bi cell. After depositing the Cr layer, the same process of rotating and deposition is repeated for Sb. Figures 1(b) and (c) show a schematic illustration of the Bi and Cr step, respectively. Then, the substrate temperature is increased to TannealT_{\mathrm{anneal}} (175°C) during which Te is continuously evaporated from the same direction as the previous elements, forming the tetradymite crystal structure. Finally, Te deposition is stopped and the substrate is cooled back down to TgrowthT_{\mathrm{growth}}. In order to achieve a certain desired film thickness, this cycle can be repeated any number of times fully automatically. Increasing the thickness per cycle, a quantity that is introduced correspondingly to a conventional growth rate and measured in quintuple layers (QLs), from 1 QL to 1.25 QL leads to a better surface morphology but decreases the crystal quality. A low annealing temperature enhances crystal quality but also increases the formation of parasitic phases on the mask and misoriented grains. Careful balancing of these two parameters leads to the growth of high quality (CrxBiySbz)2Te3(\mathrm{Cr}_{x}\mathrm{Bi}_{y}\mathrm{Sb}_{z})_{2}\mathrm{Te}_{3} (CBST) thin films. In the following, xx will be referred to as the Cr content and represents the fraction of cation sites occupied by Cr, substituting Bi and Sb.

Illustration of the digital alloying process

Figure 1. Illustration of the digital alloying process. (a) Substrate temperature versus time diagram showing the order in which the elements are deposited. (b) Schematic illustration of the Bi deposition step. The stencil mask is designed to interrupt the film in certain areas by the shadow of the mask and by that create arbitrary shapes. (c) Subsequent Cr deposition step. The substrate is rotated to a specific azimuthal angle φCr\varphi_{\mathrm{Cr}} so that Cr is deposited from the same direction as Bi previously. Thus, the angle of each cell with respect to the substrate is identical. For clarity, in the figure the cells are rotated for illustrative purpose, whereas in our actual setup the substrate is rotated while the cells are stationary.

Stencil mask fabrication and Hall bar contacting

The stencil mask layout to fabricate the CBST Hall bars is shown in Figure 2(a). The false-colored scanning electron microscopy (SEM) image shows the stencil on top of the Si(111) substrate. The Si₃N₄ layer which defines the shape of the devices is depicted in green and the supporting SiO₂ pillars in light blue. Figure 2(c) provides a closer look at the mask. It is evident that the stencil mask consists of a series of interconnected Si₃N₄ bridges, situated between the SiO₂ pillars. This allows to fully remove the Si₃N₄ layer using a first contact polymer [26] and therefore facilitates contacting the Hall bars and enables the deposition of a top gate. Figure 2(b) shows a false-colored SEM image of a Hall bar device after digital alloying growth and subsequent mask removal. During growth, all cells were successively aligned with the direction indicated by the purple arrow in Figure 2(a). The CBST thin film in the shape of a Hall bar (highlighted in purple) is through this in-situ process fully isolated from the global layer of CBST outside of the mask (grey). A capping layer of 5 nm of Al₂O₃ is

Hall bar fabrication

Figure 2. Hall bar fabrication. (a) False colored SEM image of the stencil mask under an angle on the Si(111) growth substrate. The Si₃N₄ layer which defines the shape of the devices is depicted in green and the supporting SiO₂ pillars in light blue. The purple arrow indicates how the substrate was aligned for the digital alloying step so that the MBE deposition of all materials happens from a fixed direction in alignment with the mask and under a polar angle of 42°. (b) False colored SEM image of a Hall bar device after MBE growth. After growth the Si₃N₄ layer is removed with a first contact polymer [26] only the SiO₂ pillars remain. The isolated CBST film in Hall bar shape is depicted in purple. (c) Zoom into the image from (a). The mask is made out of multiple free standing Si₃N₄ bridges in between the SiO₂ pillars to enable the mask removal via the first contact polymer [26]. (d) Zoom into the image from (b). The surface morphology of the CBST film becomes visible. (e) Optical microscopy image of the finished device after global etching of excess CBST, contact deposition (grey) and top gate deposition (gold).

Set B

deposited in-situ from all four cardinal directions so that the Hall bar is passivated. Figure 2(d) shows a close-up of the Hall bar device and reveals the surface morphology of the CBST film. In order to measure the Hall bar devices, the excess global layer of CBST is etched away while protecting the Hall bars with a photoresist and afterwards metal contacts and a top gate are deposited via lift-off processes. Figure 2(e) shows an optical microscopy image of a final device. Two main sets of Hall bar devices from (CrxBiySbz)2Te3(\mathrm{Cr}_x\mathrm{Bi}_y\mathrm{Sb}_z)_2\mathrm{Te}_3 are produced: Set A consists out of Hall bars with a width of W=3 μmW = 3\,\mu\mathrm{m} and varying Cr contents of x=0.019x = 0.019, 0.0320.032, 0.0510.051, and 0.060.06. Set B Hall bars have a Cr content of 0.0510.051 and widths including W=3 μmW = 3\,\mu\mathrm{m}, 1 μm1\,\mu\mathrm{m}, 500 nm500\,\mathrm{nm}, 300 nm300\,\mathrm{nm} and 100 nm100\,\mathrm{nm}. The Bi to Sb ratio is always roughly 1:1. The Cr contents were determined via Rutherford backscattering spectroscopy (RBS). The exact composition for each Cr content can be found in the RBS reports [27].

Structural analysis

To investigate the crystal structure, transmission electron microscopy (TEM) was used to analyse some comparable samples that were grown in the same way. Figure 3(a) shows the energy-dispersive x-ray (EDX) spectroscopy profile of a lamella cut from a 1-μm\mu\mathrm{m}-wide Hall bar made of (Cr0.03Bi0.47Sb0.5)2Te3(\mathrm{Cr}_{0.03}\mathrm{Bi}_{0.47}\mathrm{Sb}_{0.5})_2\mathrm{Te}_3. The TEM image on the xx-axis serves as a reference where the EDX profile was taken and reveals the quintuple layer structure of the MTI film. To check whether a narrower mask changes how many magnetic dopants are incorporated into the thin films, the lamella of the 1-μm\mu\mathrm{m}-wide device is compared with another lamella of a 5-μm\mu\mathrm{m}-wide one at the same composition. Figure 3(b) shows the EDX spectra of both lamella in the energy region where the Cr peaks are expected. It becomes clear that the peak height is very similar and therefore it is not expected that a more narrow masks prevents Cr atoms from reaching the devices. The most noticeable difference between the two lamellas is that the layer of the 5-μm\mu\mathrm{m}-wide Hall bar has a thickness of 14 nm while the 1-μm\mu\mathrm{m}-wide structure has a smaller film thickness of 12.8 nm. The thicknesses are determined from cross-sectional TEM images.

EDX profile and spectra of Hall bars with widths of 5 $\mu$m and 1 $\mu$m

Figure 3. Structural analysis of a previously grown sample. (a) EDX profile of a lamella cut from a Hall bar with a width of W=1 μmW = 1\,\mu\mathrm{m}. The TEM image on the xx-axis shows the quintuple layer structure clearly. (b) Comparison between EDX spectra of Hall bars with widths W=5 μmW = 5\,\mu\mathrm{m} and 1 μm1\,\mu\mathrm{m} in the energy region where the Cr peaks are expected. No qualitative difference between the spectra is found.

Magnetotransport Measurements

To investigate the low-temperature magnetotransport properties, all devices from sets A and B are measured in a variable temperature insert (VTI) cryostat with a base temperature of 1.5 K. First, the temperature-dependent (gate-dependent) magnetotransport measurements and analysis are presented for an exemplary sample with a specific Cr content of 0.051 (0.032) and a width of W=3 μmW = 3\,\mu\mathrm{m} (W=1 μmW = 1\,\mu\mathrm{m}). To study the effect of different Cr contents, the magnetotransport on Hall bars with the same dimensions but varying Cr contents are compared (set A). Further on, for a fixed Cr content the transport properties of Hall bars of different width are studied (set B).

Temperature dependent effects

First, we investigate temperature related effects by measuring the longitudinal RxxR_{xx} and Hall resistance RxyR_{xy} for all Hall bars in a wide range of temperatures. Figure 4 shows such a set of measurements at the example of a Hall bar with a width of W=3W = 3 µm and a Cr content of 0.051. In Figure 4(a) the longitudinal resistance curves over the complete magnetic field range of ±10\pm10 T is shown for temperatures from 1.5 to 14 K. One finds that RxxR_{xx} decreases with increasing temperature, indicating insulator behaviour. The trend of the zero field value of RxxR_{xx} against temperature is shown in Figure 4(d). Such a temperature dependence has been previously reported for strongly disordered CBST films, where the Fermi energy is located within the bulk band gap [28]. Figure 4(b) reveals the hysteretic nature of the longitudinal resistance for opposite directions of magnetic field sweeps.

Temperature dependent magnetotransport measurements

Figure 4. Temperature dependent magnetotransport of a 3-µm-wide Hall bar with a Cr content of 0.051. (a) RxxR_{xx} over the complete magnetic field range for temperatures ranging from T=1.5T = 1.5 K to 14 K. (b) RxxR_{xx} for different temperatures in the low magnetic field range. A hysteresis is visible between the up-sweep (dashed line) and down-sweep (solid line) of the magnetic field. (c) RxyR_{xy} for varying temperatures in a magnetic field range of ±10\pm10 T. At low fields (inset) the hysteresis is clearly visible. d) Temperature dependence of RxxR_{xx} at B=0B = 0. The decreasing trend implicates semiconducting behaviour. e) Coercive field BcB_c against temperature. The coercive field is extracted from Rxy(Bc)=0R_{xy}(B_c) = 0. The decrease of BcB_c for increasing TT is expected for ferromagnetic materials.

RxxR_{xx} exhibits a peak at the negative and positive coercive fields ±Bc\pm B_c, which decreases with temperature. This becomes more apparent from the Hall resistance data shown in Figure 4(c). The inset clearly shows how the hysteretic part of RxyR_{xy} slowly disappears at higher temperatures. The coercive field determined from Rxy(±Bc)=0R_{xy}(\pm B_c) = 0 is plotted as a function of temperature in Figure 4(e). Between 8 and 10 K BcB_c vanishes, which marks the transition from the ferromagnetic to the paramagnetic state, assuming that RxyR_{xy} is proportional to the magnetization MM of the sample. In reality, coercivity is governed by magnetization-reversal mechanisms, including domain-wall pinning, magnetic anisotropy, and thermal activation, rather than solely by the existence of ferromagnetic order [29, 30, 31]. Therefore, BcB_c vanishing can not be directly associated with the transition from ferromagnetic to paramagnetic behaviour in the magnetization MM. Still, the loss of coercivity in RxyR_{xy} can be qualitatively discussed since we are interested in the magneto-transport properties of our MTI thin films. Above 10 K the sample is behaving paramagnetic in RxyR_{xy}, which is apparent from the non-linearity at low magnetic fields in Figure 4(c). At high temperatures the shape of RxyR_{xy} at high magnetic fields is largely affected by approaching the magnetization of the layer. Therefore, it is not possible to extract the projected two-dimensional charge carrier density n2dn_{2d} at elevated temperatures from the slope of RxyR_{xy} [32, 33]. Later on, n2dn_{2d} will be discussed only from measurements taken at sufficiently low temperatures at 1.5 K, where the magnetization is saturated already. Additionally, the magnitude of the anomalous Hall effect is far from being quantized at h/e2≈25.8h/e^2 \approx25.8 kΩ\Omega. This is a result of bulk and surface transport dominating over the chiral edge channel. This together with the temperature dependence of RxxR_{xx} hints towards strong local fluctuations in the size of the magnetic gap as well as the presence of charge puddles which both cause the contribution of the chiral edge channel to be small compared to surface and bulk transport. [28, 34, 35].

Gate voltage dependent effects

Next, the effect of a gate voltage on the magnetotransport properties of a Hall bar from set B with a width of 1 μ\mum and a Cr content of 0.032 is studied. Figure 5(a) and (Figure 5b) show RxxR_{xx} as a function of magnetic field for different gate voltages VGV_G. The measurements were taken at the base temperature of 1.5 K. The longitudinal resistance at zero field changes monotonically as a function of VGV_G, which is also depicted more explicitly in Figure 5(d). Since positive gate voltages cause a decrease of Rxx(B=0)R_{xx}(B = 0) whereas negative voltages lead to a higher zero field resistance, the main charge carrier type contributing to transport can be assumed to be electrons. This assumption is consistent with the slope of the Hall voltage in Figure 5 (c) at large magnetic fields being negative. Another noticeable feature is the modulations of RxyR_{xy} as a function of BB for higher magnetic fields, as well as a relatively small change of the ordinary Hall slope for varying gate voltages. Together with the generally low value of RxyR_{xy} they again hint towards high charge disorder and the presence of electron and hole charge puddles [35, 36], in contrast to a compensation of pp- and nn-doping which would lead to a large value of RxyR_{xy}. A similar behaviour was already observed for conventionally grown CBST films on silicon substrates in our previous work [15]. For non-magnetic materials, a two-channel model can be employed to extract the charge carrier densities for electrons and holes separately [37], but in our case the presence of the ferromagnetic hysteresis makes this impossible [38]. Therefore, the charge carrier densities extracted from a regular Hall analysis might not produce the true value. The coercive field is mostly constant as a function of VGV_G, as shown in Figure 5(e). The error bars stem from determining the positive and negative coercive field and averaging them. Gate voltage independent ferromagnetic properties, such as BcB_c, hint towards the charge carrier independent Van Vleck mechanism being mainly responsible for ferromagnetic order in our films [12, 13].

Gate voltage dependent magnetotransport measurements

Figure 5. Gate voltage dependent magnetotransport measurements of a 1-μ\mum-wide Hall bar with a Cr content of 0.032. (a) and (b) Longitudinal magnetoresistance RxxR_{xx} for gate voltages between VG=−10V_G = -10 V and 10 V over a large range of BB-field and in a small range around zero field, respectively. (c) RxyR_{xy} for different gate voltages. The inset shows a zoom-in of the hysteresis. (d) The zero field resistance Rxx(B=0)R_{xx}(B = 0) behaves monotonically as a function of gate voltage. (e) The coercive field BcB_c as a function of VGV_G stays basically constant.

Effect of chromium content

Next, we investigate the effect of different Cr contents. Therefore, set A consisting of 3-µm-wide Hall bars with x=0.019x = 0.019, 0.032, 0.051, and 0.6 are compared. The raw data of all set A samples can be found in the Supplemental Material [39]. Figure 6(a) shows the BcB_c vs. TT dependence at each Cr content, respectively. It becomes clear that the coercive field becomes zero at lower temperatures for lower values of xx. Figure. Figure 6(b) shows BcB_c at base temperature plotted against the Cr content and reveals a linear increase. The results given in Figure 6(a) and (b) are expected from a classical ferromagnetism standpoint and consistent with previous studies on the same material [12, 13, 14, 40]. In Figure 6(c) the electron concentration n2dn_{2d} at 1.5 K is plotted as a function of Cr content. We found in our previous study [15] that Cr acts as a pp-type dopant in Cr-doped (Bi1−ySby)2Te3(\mathrm{Bi}_{1-y}\mathrm{Sb}_y)_2\mathrm{Te}_3. The same result was also found by other authors [41]. In contrast, in Figure 6(c) we find the opposite behaviour where n2dn_{2d} increases with increasing Cr content. A possible explanation is the formation of charge puddles where similar to our study on conventionally MBE-grown CBST thin films the classical Hall analysis breaks down and gives misleading results due to the compensation of electrons and holes [15].

Comparison of relevant transport parameters of 3-µm-wide Hall bars for various Cr contents

Figure 6. Comparison of relevant transport parameters of 3-µm-wide Hall bars for various Cr contents xx. (a) BcB_c versus TT for Cr contents from x=0.019x = 0.019 to 0.060.06. (b) BcB_c at 1.5 K against Cr content. (c) Electron concentration at n2dn_{2d} at 1.5 K as a function of xx.

Effect of Hall bar dimensions

Finally, set B Hall bars with a fixed Cr content of 0.051 and widths varying from W=3 μmW = 3\,\mu\mathrm{m} down to 100 nm are compared. Figure 7(a) shows a basically constant BcB_c with a small trend of lower values towards smaller Hall bar widths. This trend is possibly due to the thickness reduction of narrower Hall bars found from the TEM analysis. As can be seen in Figure 7(b), the zero-field resistance of the Hall bars Rxx(B=0)R_{xx}(B = 0) decreases with increasing width WW. This decrease follows a 1/W1/W dependence (dashed black line), indicating a basically constant resistivity, independent on the sample width. This is also in line with the electron concentration n2dn_{2d} shown in Figure 7(c), which was extracted from the slope of the Hall signal at large magnetic fields, i.e. where the magnetization due to the Cr-dopants is saturated [32]. One finds that n2dn_{2d} remains in range of about 1×10141 \times10^{14} to 5×1014 cm−25 \times10^{14}\ \mathrm{cm}^{-2}. The value itself seems to fluctuate a bit from device to device, which can be expected for strongly disordered thin films.

Comparison of Hall bars with $x = 0.051$ Cr content and varying widths $W$

Figure 7. Comparison of Hall bars with x=0.051x = 0.051 Cr content and varying widths WW. (a) BcB_c as a function of Hall bar width WW. (b) Zero field resistance Rxx(B=0)R_{xx}(B = 0) as a function of WW. The black dashed line represents the expected 1/W1/W dependence of Rxx(B=0)R_{xx}(B = 0) (c) Electron concentration n2dn_{2d} as a function of WW.

Conclusion

We successfully fabricated micro- and nanometer-sized Hall bars from Cr-doped (BiySb1−y)2Te3(\mathrm{Bi}_{y}\mathrm{Sb}_{1-y})_{2}\mathrm{Te}_{3} using digital alloying and stencil-mask techniques. A structural analysis of two Hall bar devices via TEM and EDX confirmed this through the crystal structure and the incorporated materials. The analysis also eliminates the suspicion that the stencil-mask hinders the deposition of materials for narrow structures. Subsequently, Hall bars with different Cr contents and a range of different widths are investigated in magnetotransport measurements. First, the temperature dependence of a specific Hall bar is discussed and implications of disorder and charge puddles are found from the temperature dependence of RxxR_{xx}. Furthermore, the transition from ferromagnetic to paramagnetic behaviour in RxyR_{xy} with increasing temperature is observed and discussed with respect to the difference between coercivity in a Hall measurement and general ferromagnetic order. At the example of another Hall bar, gate voltage dependent measurements are presented and further implications for a disordered potential landscape are found. This is supported by contradictory trends of n2dn_{2d} for increasing Cr contents. The coercive field and the transition temperature from ferromagnetic to paramagnetic transport signatures behave as expected if more Cr atoms are incorporated into the crystal. For a fixed Cr content, no significant width dependence of the coercive field is found and the sheet resistance RxxR_{xx} follows the expected 1/W1/W trend. The electron concentration remains in the same order of magnitude and is generally not expected to be dependent on the Hall bar dimensions. Some fluctuations in the value can be again attributed to disorder.

Acknowledgments

We thank Herbert Kertz for technical assistance. This work is supported by the QuantERA grant MAGMA and by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Grant No. 491798118. In addition, we received funding from DFG under Germany’s Excellence Strategy - Cluster of Excellence Matter and Light for Quantum Computing (ML4Q) EXC 2004/2 – 39054736 as well as financial support by the Bavarian Ministry of Economic Affairs, Regional Development and Energy within Bavaria’s High-Tech Agenda Project ”Bausteine für das Quantencomputing auf Basis topologischer Materialien mit experimentellen und theoretischen Ansätzen” (grant no. 07 02/686 58/1/21 1/22 2/23). J.K., M.K., and T.S. would like to express their gratitude for the support provided as part of the ASPIRE project JPMJAP2338.

Data Availability

The data that support the findings of this article are openly available [27].

References

Supplemental Material: Digitally-alloyed magnetic topological insulator nanostructures grown via stencil mask molecular beam epitaxy

Surface Morphology

Supplemental Figure S1 shows scanning electron microscopy (SEM) images of four Cr-doped (Bi Sb)2Te3(\mathrm{Bi}\ \mathrm{Sb})_{2}\mathrm{Te}_{3} (CBST) thin films grown via digital alloying with varying Cr contents of 0.019, 0.032, 0.051, and 0.06. From these SEM images one finds that the surface becomes slightly rougher for increasing Cr content, which is in line with our previous study on CBST thin films with different Cr contents [S1].

SEM images of CBST thin films at Cr contents 0.019, 0.032, 0.051, and 0.06

Figure S1. SEM images of a CBST thin films at each Cr content that is discussed in the main text.

Raw Data of Samples

Set A

Supplemental Figure S2 presents the raw data of the samples that are referred to as set A in the main text. Set A contains Hall bars with a width of 3 μm3\,\mu\mathrm{m} and Cr contents of 0.019, 0.032, 0.051, and 0.06. All devices are measured for temperatures from 1.5 K up to at least 10 K or higher over a large magnetic field range between -10 T to 10 T. For each device RxxR_{xx} shows a decreasing trend with increasing temperature like the device shown in the main text. From the curve of RxyR_{xy} we extract the coercive field BcB_{c} as a function of temperature for each Cr content as well as the charge carrier density at 1.5 K.

Raw data of set A for Hall bars with different Cr contents

Figure S2. Raw data of set A. Hall bars with a width of 3 μm3\,\mu\mathrm{m} are measured for different Cr contents ranging from 0.019 to 0.06.

Set B

Set B contains Hall bars with a Cr content of 0.051 and widths including 3 µm, 1 µm, 500 nm, 300 nm, and 100 nm. Supplemental Figure S3 once again shows the raw data of RxxR_{xx} and RxyR_{xy} for all Hall bars in this set. The Hall bar with a width of 3 µm and 0.051 Cr content is not shown again as it is already included in the data for group A. Supplemental Figure S3(b) shows the curve for different gate voltages since high field data for this device was not recorded for different temperatures. From set B samples we extract the charge carrier density at 1.5 K (at zero gate voltage) at high magnetic fields above 3 T. In the sample shown in Supplemental Figure S3(b), RxxR_{xx} at a constant field also does not show a clear gate dependence as the sample from the main text. This could be due to charge puddles [S2–S4], i.e. electron and hole contributions, being affected differently by a gate voltage, causing RxyR_{xy} to change non-monotonically.

References

Raw data of set B for Hall bars with a Cr content of 0.051 measured for different Hall bar widths

Figure S3. Raw data of set B. Hall bars with a Cr content of 0.051 are measured for different Hall bar widths WW.

References

  1. [1]C.-Z. Chen, Y.-M. Xie, J. Liu, P. A. Lee, and K. T. Law, Quasi-one-dimensional quantum anomalous Hall systems as new platforms for scalable topological quantum computation, Physical Review B 97, 104504 (2018).
  2. [2]X.-L. Qi, T. L. Hughes, and S.-C. Zhang, Chiral topological superconductor from the quantum Hall state, Physical Review B 82, 184516 (2010).DOI
  3. [3]Y. Tokura, K. Yasu da, and A. Tsukazaki, Magnetic topological insulators, Nature Reviews Physics 1, 126 (2019).DOI
  4. [4]D. Burke, D. Heffels, K. Moors, P. Schüffelgen, D. Grützmacher, and M. R. Connolly, Robust Majorana bound states in magnetic topological insulator nanoribbons with fragile chiral edge channels, Physical Review B 109, 045138 (2024).DOI
  5. [5]E. Zsurka, D. Di Miceli, J. Legendre, L. Serra, D. Grützmacher, T. L. Schmidt, and K. Moors, Optimizing proximitized magnetic topological insulator nanoribbons for Majorana bound states, Physical Review B 112, 094504 (2025).DOI
  6. [6]T. Jansen, E. Kochetkova, A. Isaeva, A. Brinkman, and C. Li, Josephson coupling across magnetic topological insulator MnBi₂Te₄, Communications Materials 5, 214 (2024).DOI
  7. [7]E. Zhang, Z.-T. Sun, Z. Jia, J. Yang, J. Yan, L. Ai, Y.-M. Xie, Y. Zhang, X.-J. Gao, X. Xu, S. Liu, Q. Ma, C. Hu, X. Kou, J. Zou, N. Ni, K. T. Law, S. Dong, and F. Xiu, Observation of edge supercurrent in topological antiferromagnet MnBi₂Te₄-based Josephson junctions, Science Advances 11, eads8730 (2025).DOI
  8. [8]A. Uday, G. Lippertz, K. Moors, H. F. Legg, R. Joris, A. Bliesener, L. M. C. Pereira, A. A. Taskin, and Y. Ando, Induced superconducting correlations in a quantum anomalous Hall insulator, Nature Physics 20, 1589 (2024).
  9. [9]P. Mandal, N. Taufertshöfer, L. Lunczer, M. P. Stehno, C. Gould, and L. W. Molenkamp, Finite field transport response of a dilute magnetic topological insulator-based Josephson junction, Nano Letters 22, 3557 (2022).DOI
  10. [10]J. Legendre, E. Zsurka, D. Di Miceli, L. Serra, K. Moors, and T. L. Schmidt, Topological properties of finite-size heterostructures of magnetic topological insulators and superconductors, Physical Review B 110, 075426 (2024).DOI
  11. [11]K. Moors and G. Yin, Robust quantum anomalous Hall effect with spatially uncorrelated disorder, Physical Review B 112, 224437 (2025).DOI
  12. [12]X. Kou, M. Lang, Y. Fan, Y. Jiang, T. Nie, J. Zhang, W. Jiang, Y. Wang, Y. Yao, L. He, and K. L. Wang, Interplay between different magnetisms in Cr-doped topological insulators, ACS Nano 7, 9205 (2013).DOI
  13. [13]C. Chang, J. Zhang, M. Liu, Z. Zhang, X. Feng, K. Li, L. Wang, X. Chen, X. Dai, Z. Fang, X. Qi, S. Zhang, Y. Wang, K. He, X. Ma, and Q. Xue, Thin films of magnetically doped topological insulator with carrier-independent long-range ferromagnetic order, Advanced Materials 25, 1065 (2013).DOI
  14. [14]Y. Ou, C. Liu, L. Zhang, Y. Feng, G. Jiang, D. Zhao, Y. Zang, Q. Zhang, L. Gu, Y. Wang, K. He, X. Ma, and Q.-K. Xue, Heavily Cr-doped (Bi,Sb)₂Te₃ as a ferromagnetic insulator with electrically tunable conductivity, APL Materials 4, 086101 (2016).DOI
  15. [15]J. Karthein, J. Buchhorn, K. Underwood, A. R. Jalil, M. Vaßen-Carl, P. Schüffelgen, D. Grützmacher, and T. Schäpers, Controlling the magnetotransport properties of magnetic topological insulator Crₓ(BiᵧSb₁−ᵧ)₂−ₓTe₃ thin films via molecular beam epitaxy, Physical Review Materials 9, 124202 (2025).
  16. [16]P. Schüffelgen, D. Rosenbach, C. Li, T. W. Schmitt, M. Schlewevoigt, A. R. Jalil, S. Schmitt, J. Kölzer, M. Wang, B. Bennemann, U. Parlak, L. Kibkalo, S. Trellenkamp, T. Grap, D. Mertens, M. Luysberg, G. Mussler, E. Berenschot, N. Tas, A. A. Golubov, A. Brinkman, T. Schäpers, and D. Grützmacher, Selective area growth and stencil lithography for in situ fabricated quantum devices, Nature Nanotechnology 14, 825 (2019).DOI
  17. [17]G. Qiu, P. Zhang, P. Deng, S. K. Chong, L. Tai, C. Eckberg, and K. L. Wang, Mesoscopic transport of quantum anomalous Hall effect in the submicron size regime, Physical Review Letters 128, 217704 (2022).DOI
  18. [18]M. P. Andersen, L. K. Rodenbach, T. T. Rosen, S. C. Lin, L. Pan, P. Zhang, L. Tai, K. L. Wang, M. A. Kastner, and D. Goldhaber-Gordon, Low-damage electron beam lithography for nanostructures on Bi₂Te₃-class topological insulator thin films, Journal of Ap- I. T. Rosen, C. J. Trimble, M. P. Andersen, E. Mikheev, Y. Li, Y. Liu, L. Tai, P. Zhang, K. L. Wang, Y. Cui, M. A. Kastner, J. R. Williams, and D. Goldhaber-Gordon, Fractional action at Josephson effect in a topological insulator proximitized by a self-formed superconductor, Physical Review B 110, 064511 (2024).
  19. [20]R. C. Miller, A. C. Gossard, D. A. Kleinman, and O. Munteanu, Parabolic quantum wells with the GaAs-AlₓGa₁₋ₓAs system, Phys. Rev. B 29, 3740(R) (1984).DOI
  20. [21]J. G. Cody, D. L. Mathine, R. Droopad, G. N. Maracas, R. Rajesh, and R. W. Carpenter, Application of the digital alloy composition grading technique to strained InGaAs/GaAs/AlGaAs diode laser active regions, Journal of Vacuum Science & Technology B 12, 1075 (1994).DOI
  21. [22]R. Kaspi and G. P. Donati, Digital alloy growth in mixed As/Sb heterostructures, Journal of Crystal Growth 251, 515 (2003).DOI
  22. [23]A. Taylor, C. Mortensen, R. Rostek, N. Nguyen, and D. C. Johnson, Vapor annealing as a post-processing technique to control carrier concentrations of bi2te3 thin films, Journal of Electronic Materials 39, 1981 (2009).DOI
  23. [24]N. Peranio, M. Winkler, D. Bessas, Z. Aabdin, J. König, H. Böttner, R. Hermann, and O. Eibl, Room-temperature mbe deposition, thermoelectric properties, and advanced structural characterization of binary bi2te3 and sb2te3 thin films, Journal of Alloys and Compounds 521, 163 (2012).DOI
  24. [25]Y. Gong, J. Guo, J. Li, K. Zhu, M. Liao, X. Liu, Q. Zhang, L. Gu, L. Tang, X. Feng, D. Zhang, W. Li, C. Song, L. Wang, P. Yu, X. Chen, Y. Wang, H. Yao, W. Duan, Y. Xu, S.-C. Zhang, X. Ma, Q.-K. Xue, and K. He, Experimental realization of an intrinsic magnetic topological insulator*, Chinese Physics Letters 36, 076801 (2019).DOI
  25. [26]Photonic Cleaning Technologies, First Contact™ Polymer, https://www.photoniccleaning.com/, commercial polymer cleaner for precision optics.
  26. [27]Karthein, Jan, Vaßen-Carl, Max, Maier, Lars, Jalil, Abdur Rehman, Toehgiono, Gion, Thieme, Josua, Feldewerth, Julian, Bennemann, Benjamin, Domnick, Joscha, Kawano, Taizo, Otsubo, Sosuke, Heller, René, Ishihara, Jun, Grützmacher, Detlev, Kohda, Makoto, Schüffelgen, Peter, and Schäpers, Thomas, Data for: Digitally-alloyed magnetic topological insulator nanostructures grown via stencil mask molecular beam epitaxy, JülichDATA (2026).arxiv.org/abs/2611.18280
  27. [28]Q. Guo, Y. Wu, D. Wang, G. Han, X. Wang, L. Fu, L. Wang, W. He, T. Zhu, Z. Zhu, T. Liu, G. Yu, and J. Teng, Enhanced negative magnetoresistance near the charge neutral point in Cr doped topological insulator, RSC Advances 11, 13964 (2021).DOI
  28. [29]K. J. Strnat, H. F. Mildrum, M. Tokunaga, and H. Harada, Evidence for domain wall pinning by a magnetic grain-boundary phase in sintered Nd-Fe-B based permanent magnets, Journal of Applied Physics 63, 3321 (1988).DOI
  29. [30]G. Vértesy, I. Tomás, and Z. Vértesy, On the temperature dependence of domain wall pinning field in soft, uniaxial magnetic materials, Journal of Physics D: Applied Physics 35, 625 (2002).DOI
  30. [31]C.-b. Rong, H.-w. Zhang, B.-g. Shen, and J. P. Liu, Mechanism of the anomalous temperature dependence of coercivity in Sm(Co,Fe,Cu,Zr)ₓ high-temperature magnets, Applied Physics Letters 88, 042504 (2006).DOI
  31. [32]H. Ohno, Making nonmagnetic semiconductors ferromagnetic, Science 281, 951 (1998).pubmed.ncbi.nlm.nih.gov/9703503
  32. [33]J. J. Berry, S. J. Potashnik, S. H. Chun, K. C. Ku, P. Schiffer, and N. Samarth, Two-carrier transport in epitaxially grown MnAs, Physical Review B 64, 052408 (2001).DOI
  33. [34]Y. X. Chong, X. Liu, R. Sharma, A. Kostin, G. Gu, K. Fujita, J. C. S. Davis, and P. O. Sprau, Severe dirac mass gap suppression in Sb₂Te₃-based quantum anomalous Hall materials, Nano Letters 20, 8001 (2020).DOI
  34. [35]G. Lippertz, A. Bliesener, A. Uday, L. M. C. Pereira, A. A. Taskin, and Y. Ando, Current-induced breakdown of the quantum anomalous Hall effect, Physical Review B 106, 045419 (2022).DOI
  35. [36]G. Qiu, H.-Y. Yang, S. K. Chong, Y. Cheng, L. Tai, and K. L. Wang, Manipulating topological phases in magnetic topological insulators, Nanomaterials 13, 2655 (2023).DOI
  36. [37]N. Ito, R. Matsutomi, and T. Okamoto, Cancellation of electron and hole contributions to the Hall effect in ultrathin Bi films grown on GaAs(110), Physical Review B 105, 205434 (2022).DOI
  37. [38]Y. He, J. Gayles, M. Yao, T. Helm, T. Reimann, V. N. Strocov, W. Schnelle, M. Nicklas, Y. Sun, G. H. Fecher, and C. Felser, Large linear non-saturating magnetoresistance and high mobility in ferromagnetic MnBi, Nature Communications 12, 4576 (2021).DOI
  38. [39]See Supplemental Material for information on surface morphology and raw data of all measured samples.
  39. [40]M. Ye, W. Li, S. Zhu, Y. Takeda, Y. Saitoh, J. Wang, H. Pan, M. Nurmamat, K. Sumida, F. Ji, Z. Liu, H. Yang, Z. Liu, D. Shen, A. Kimura, S. Qiao, and X. Xie, Carrier-mediated ferromagnetism in the magnetic topological insulator Cr-doped (Sb,Bi)₂Te₃, Nature Communications 6, 8913 (2015).
  40. [41]P. Gultom, C.-C. Hsu, M. K. Lee, S. H. Su, and J.-C.-A. Huang, Epitaxial growth and characterization of nanoscale magnetic topological insulators: Cr-doped (Bi₀.₄Sb₀.₆)₂Te₃, Nanomaterials 14, 157 (2024).DOI

Paper details

vlnc:W_4gkv6aeg
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents