† Corresponding author. E-mail:
Project supported by the National Basic Research Program of China (Grants Nos. 2015CB921400 and 2011CB921802) and the National Natural Science Foundation of China (Grants Nos. 11374057, 11434003, and 11421404).
The surface impurity effect on the surface-state conductivity and weak antilocalization (WAL) effect has been investigated in epitaxial Bi (111) films by magnetotransport measurements at low temperatures. The surface-state conductivity is significantly reduced by the surface impurities of Cu, Fe, and Co. The magnetotransport data demonstrate that the observed WAL is robust against deposition of nonmagnetic impurities, but it is quenched by the deposition of magnetic impurities which break the time reversal symmetry. Our results help to shed light on the effect of surface impurities on the electron and spin transport properties of a 2D surface electron systems.
Spin-polarized two-dimensional (2D) surface states arising from the Bi or Bi-based topological insulators are of large potential interest for the next generation of magnetic memory and logic devices.[1–3] Recently in particular, it has been shown that the Rashba–Edelstein effect and the inverse Rashba–Edelstein effect can be regarded as a counterpart of the spin Hall effect and inverse spin Hall effect for spin-charge interconversion.[4–8] When an in-plane current flows through a surface state, due to the spin momentum locking nature resulted from the Rashba effect, the charge flow is accompanied by a non-equilibrium spin polarization with direction perpendicular to both the current direction and the surface normal. By coupling this spin polarized current to an adjacent magnetic layer, an efficient spin–orbit induced torque will act on the magnet as shown by Mellnik et al. in a bilayer system Bi2Se3/permalloy.[6] By using spin pumping, Rojas Sanchez et al. successfully injected a spin current from an NiFe layer into a Bi/Ag bilayer and detected the resulting charge current in the 2D interface. Usually, these kinds of devices consist of a thin film that host the 2D surface states and a magnetic layer. The direct contact between these two would introduce impurities into the 2D surface states, which are in general very sensitive to surface adsorption atoms. Thus, it is important to know how the magnetic or nonmagnetic impurities that come from the magnetic layer would affect the electron and spin transport in the 2D surface electron systems.
Angle-resolved photoemission spectroscopy (ARPES) measurements have shown that the surface states of Bi are highly metallic in contrast to the semimetallic nature of bulk Bi.[9–11] Especially, by reducing thickness, the Bi film interior can be transformed into semiconductor.[12–14] These make the strong spin splitting surface state of Bi ultrathin film one of the most promising platforms for application. By combining scanning tunneling microscopy with density functional theory, Klein et al. has shown that the Bi (111) surface provides a well-defined incorporation site in the first bilayer that traps highly coordinating atoms such as transition metals or noble metals.[15] However, the transport study of the impurities effect on surface states are mainly in poly crystalline Bi films,[16–18] in which the surface has no well defined orientation. It is still not clear how the impurities would affect the surface conductivity and the phase coherence length in the Bi (111) surface states. On the other hand, as for application in real devices, it is also important to know whether the surface states are robust in air conditions, and show up in the physical properties of the Bi (111) films.
In this work, we introduce magnetic (Fe, Co) and nonmagnetic (Cu) impurities onto the surface of single-crystalline Bi (111) films and performed magnetotransport measurements. A significant reduction of surface conductivity is found for 0.5 monolayer (ML) of Cu, Fe and Co impurities on 5-nm Bi (111) film. The magnetotransport data demonstrate that the observed WAL is robust against deposition of nonmagnetic impurities, but it is quenched by the deposition of magnetic impurities which breaks locally the time reversal symmetry. More surprisingly, the Bi (111) surface state is found to be very robust against oxidation, even with the fact that Bi is believed to be a prototypical semimetal with topologically trivial electronic band structure.[19] Our results help to shed light on the effect of surface impurities on the electron and spin transport properties of a 2D surface electron system.
The Bismuth films were grown by molecular beam epitaxy on a semi-insulating (111)-oriented Si substrate and characterized in situ using reflection high energy electron diffraction (RHEED). A clean Si (111)-7×7 surface was obtained by a cycle of e-beam heat treating as shown by the RHEED pattern in Fig.
The red curve in Figs.
We first show how the surface impurities would affect the surface conductivity. It can be clearly seen from the blue curve in Figs.
In fact, the different influence for the nonmagnetic and magnetic impurities is reasonable considering the special electronic band structure of the surface states. It has been shown by (spin- and) angle-resolved photoemission spectroscopy (ARPES) that the surfaces of Bi (111) films are characterized by a quasi-two-dimensional metallic surface states with strong spin-splitting.[9–11] Due to time reversal symmetry, the spin-conserving backscattering from +k to −k states of propagating electrons near the Fermi level EF is strongly suppressed, which is confirmed by scanning tunneling microscopy (STM) experiment.[23] Then, when the adatoms with magnetic moments are introduced on to the surface, the local time-reversal symmetry will be broken; hence the direct (spin-conserving) backscattering channel for electrons and holes is now activated. This increasing scattering probability will lead to a reduction of the electron mean free path. This is why the Fe and Co impurities on Bi (111) will lead to a more significant reduction of the surface conductivity compared to Cu impurities, which can be regarded as simple scattering centers. Additionally, no anomalous Hall effect is observed in Fe and Co deposited samples, which seems to suggest that the magnetizations of the Fe and Co clusters are random in direction.
We now concentrate on the impurity effect on the coherent transport of electrons in the surface states. The red curve in Fig.
The broadening of the magnetoresistivity curve for the Cu impurities could be understood as the reduced phase coherence length lφ induced by the additional Cu surface scattering centers (which will be discussed in detail in the following). While the magnetic impurities, which would form clusters on the Bi (111) surface, will break the locally time-reversal symmetry and will lead to a crossover from the symplectic to unitary classes.[28] The magnetoresistivity of the 5-nm sample with 0.5-ML Fe deposition shows the intermediate regime in this crossover process. For the Co impurities, the magnetoresistivity shows an almost B2 dependence on the magnetic field, which means the WAL is completely suppressed. The very different effect of magnetic and nonmagnetic impurities in our data clearly associates the WAL effect we observed with the surface states. Otherwise it should not be so sensitive to the surface modification, if the WAL originates from the Bi film bulk states. It should be mentioned that even with 1 ML of Fe impurities on top, the WAL also cannot be completely suppressed, while 0.25-ML Co impurities are enough to kill the WAL behavior. A plausible explanation for this could be that the exchange coupling in Co is much stronger than that in Fe. Nevertheless, these results demonstrate that the Bi (111) surface states are robust against nonmagnetic disorder perturbation, while quite sensitive to the surface magnetic impurities.
Next, we will go into details of how different impurities will affect the phase coherence length lφ as mentioned before. In fact, detailed knowledge in the quantum corrections to magnetoconductivity may provide an alternative technique for identifying the surface states. As for a 2D electronic system, the magnetoconductivity data can be described by the two dimensional Hikami–Larkin–Nagaoka (HLN) theory in the presence of strong spin–orbit coupling:[28]
By applying this fitting procedure to the curves at different temperature, we can extract the temperature dependence of the α and the phase coherence length lφ. The detailed results are shown in Fig.
The influence of the 0.5-ML Fe impurities is also considered here. The result is summarized in Fig.
Finally we show that the metallic surface states of Bi (111) thin films are very robust against surface oxidation. The experiment is specially designed in the following way: after an 8.0-nm thick Bi (111) thin film was epitaxially grown on Si (111), only half of the sample was covered in-situ by the 4-nm MgO capping layer, then the whole sample was taken out of the UHV chamber to be exposed in air for 1 hour at 300 K. After this process, the uncovered Bi (111) surface will be oxidized while the other should be protected because of the MgO capping layer. Then the sample was put back into the UHV chamber and the oxidized Bi (111) part was capped by 4-nm MgO. Figure
In conclusion, we show that the surface states conductivity of Bi (111) films are very sensitive to surface impurities modification. The WAL effect, which arises from the 2D strong spin orbital coupling surface states, is suppressed by magnetic impurities, but it is quite robust against nonmagnetic impurities. Even more dramatically, the surface states are very robust against the surface oxidation. We believe these unique properties could possibly make the Bi surface states useful in future spintronics devices.
1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | |
15 | |
16 | |
17 | |
18 | |
19 | |
20 | |
21 | |
22 | |
23 | |
24 | |
25 | |
26 | |
27 | |
28 |